| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Special Reports |
1 From the Department of Radiology, University of California, San Diego, 200 W Arbor Dr, San Diego, CA 92103-8756. From the 2002 RSNA Annual Meeting. Received October 9, 2004; revision requested December 21; revision received January 11, 2005; accepted February 1. Address correspondence to G.C. (e-mail: gcasola{at}ucsd.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: This HIPAA-compliant study received institutional review board approval, with waiver of informed consent. The radiologic reports of 1192 consecutive patients who underwent whole-body CT screening of the chest, abdomen, and pelvis at an outpatient imaging center from January to June 2000 were reviewed. Scans were obtained with electron-beam CT without oral or intravenous contrast material. Reported imaging findings and recommendations were retrospectively tabulated and assigned scores. Descriptive statistics were used (means, standard deviations, and percentages); comparisons between subgroups were performed with univariate analysis of variance and 
2 or Fisher exact tests.
RESULTS: Screening was performed in 1192 patients (mean age, 54 years). Sixty-five percent (774 of 1192) were men and 35% (418 of 1192) were women. Nine hundred three (76%) of 1192 patients were self referred, and 1030 (86%) of 1192 subjects had at least one abnormal finding described in the whole-body CT screening report. There were a total of 3361 findings, with a mean of 2.8 per patient. Findings were described most frequently in the spine (1065 [32%] of 3361), abdominal blood vessels (561 [17%] of 3361), lungs (461 [14%] of 3361), kidneys (353 [11%] of 3361), and liver (183 [5%] of 3361). Four hundred forty-five (37%) patients received at least one recommendation for further evaluation. The most common recommendations were for additional imaging of the lungs or the kidneys.
CONCLUSION: With whole-body CT screening, findings were detected in a large number of subjects, and most findings were benign by description and required no further evaluation. Thirty-seven percent of patients had findings that elicited recommendations for additional evaluation, but further research is required to determine the clinical importance of these findings and the effect on patient care.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Some investigators have stated that whole-body CT screening is cost effective on the basis of sparse data and controversial assumptions (1,9); other investigators have cited anecdotal experience to illustrate the potential pitfalls of CT screening, particularly the high number of false-positive test results and the consequences of them (10). To our knowledge, no cost-effectiveness analysis or large-scale randomized clinical trials have been performed for evaluation of whole-body CT screening.
Prior to planning a large-scale randomized clinical trial for whole-body CT screening, it is necessary to determine the spectrum and frequency of imaging findings and follow-up recommendations in an asymptomatic population in which screening was performed, particularly in a private imaging setting, as this setting represents the current majority of whole-body CT screening centers in the United States (5,8). Therefore, the purpose of this study was to retrospectively determine the frequency and spectrum of findings and recommendations reported with whole-body CT screening at a community screening center.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The review was performed by a research team in an academic hospital in the same city in which the imaging center was located. There was no financial relationship between the research team and the imaging center. The research team was granted approval by the institutional review board to retrospectively review reports issued by the imaging center. The institutional review board waived patient informed consent. This study was Health Insurance Portability and Accountability Act compliant.
Patients were either self referred or physician referred for whole-body CT screening; the type of referral (self vs physician) was recorded by the imaging center, but the reason for referral, if any, was not. Patients were charged approximately $1000 prior to whole-body CT screening. Third-party providers did not cover this charge. On arrival at the imaging center, patients watched a 5–10-minute instructional video produced by the imaging center, filled out a brief questionnaire in regard to medical history, and signed a consent form in the presence of the technologist (a radiologist was not present). The consent form indicated that there was a risk from radiation and that sensitivity of the examination for lesion detection was limited without contrast material. Premenopausal women were asked about the possibility of pregnancy; pregnant women and those who were uncertain about pregnancy were not scanned.
Whole-Body CT Screening Technique
Contiguous 6-mm-thick images were obtained through the chest, abdomen, and pelvis (anatomic landmarks were not known to the research team) with electron-beam CT (C-150XP; Imatron, South San Francisco, Calif). Typical parameters were 200 mAs and 130 kVp (it was unknown whether these parameters were adjusted in individual cases). Oral or intravenous contrast material was not administered. Although electron-beam CT was initially introduced for cardiac scanning, it has been widely used to image the entire body (11,12); however, its reliability for certain uses, such as bone density quantification, has not been established.
Interpretation of Findings and Radiology Reports
Standardized coronary artery calcium screening was performed at each study, and findings were interpreted by a cardiologist. The results of coronary artery calcium screening were not analyzed.
Initial interpretation of scans was performed remotely on a gray-scale monitor with a resolution of 2048 x 2560 pixels (InSight Diagnostic Imaging Workstation; Neo Imagery Technologies, City of Industry, Calif) and three-dimensional reformation capabilities by one of two board-certified radiologists who had 22 and 7 years of experience and who were contracted by the imaging center and were not part of the research team.
Dictated reports were mailed to the patient and/or referring physician. The reports were organized into three main sections: (a) patient identification information, (b) findings, and (c) impressions. The patient identification section included the patient's age, sex, and referral type. Information about medical history and the reason for referral were inconsistently provided. Patients' ethnic backgrounds were not mentioned. The findings section was subdivided into (a) heart and coronary artery calcium screening (not analyzed in this study), (b) lungs and mediastinum, (c) breasts, (d) spine, and (e) other (including abdomen and pelvis). The impressions section included summary statements and recommendations, if any, for additional evaluation. The types of findings reported, the lexicon used for description, and the type and interval of diagnostic follow-up were not standardized.
Data Collection
Dictated reports were retrospectively reviewed in consensus by a panel of three research team members: an abdominal radiologist with 25 years of experience (G.C.), a thoracic radiologist with 10 years of experience (D.L.L.), and a research fellow with 4 years of experience (C.D.F.). On the basis of the review of the dictated report, several variables were assigned scores (as listed later), and variables and scores were entered manually into a computer spreadsheet by the research fellow. Patient questionnaires, consent forms, and actual images were not available for review.
Variables Collected
Patient information.—Patient age, sex, referral type, and, if provided, data about the medical history were recorded. Disease duration, stage, and severity were not recorded (not consistently reported).
Reported imaging findings.—Imaging findings were categorized according to anatomic location. Each anatomic location was assigned a score in a binary fashion as having or not having a described imaging finding. An imaging finding was defined as any reported finding of potential clinical importance and included neoplastic or possibly neoplastic, inflammatory, infectious, degenerative, vascular, and metabolic abnormalities, as well as compression fractures and pars defects of the spine. Congenital abnormalities (eg, transitional lumbosacral vertebra, pectus excavatum), bony deformities (eg, scoliosis), postsurgical findings (eg, surgically resected tissue), healed rib fractures, and phleboliths were not included. Findings described with qualifiers such as "probable," "consistent with," "suggestive of," and "indicative of" were assigned scores as positive; findings described as "equivocal," "unlikely," or "doubtful" or those described with other similar qualifiers were assigned scores as negative. We did not record the frequency with which specific qualifiers were used.
In cases in which multiple lesions of the same type were present, multiple lesions were counted as a single finding (eg, multiple lung nodules or gallstones).
Because a standardized lexicon was not used for the dictated reports, the research panel retrospectively standardized the terminology for the description of imaging findings. The terminology used in this manuscript, therefore, is not necessarily identical to that used in the dictated reports, although the meaning was replicated as faithfully as possible. All dimensional measurements mentioned in the report were recorded. Lesion and tissue attenuation values were not routinely reported, and, thus, only subjective relative attenuation (eg, hypoattenuation, isoattenuation, hyperattenuation) was recorded, if mentioned.
Quantitative CT was used to determine bone marrow density. An impression of bone marrow density was stated by the dictating radiologists, but actual measurements were not mentioned in the report.
Reported recommendations.—The panel retrospectively recorded the number of recommendations for additional evaluation per patient and per imaging finding. The suggested type of follow-up and follow-up interval were also documented.
Statistical Analysis
The number of imaging findings and recommendations were analyzed descriptively. Means and standard deviations were reported for quantitative measures, and percentages were reported for categoric data. Patients were also stratified according to sex and age group (<40, 40–49, 50–59, 60–69, and 
70 years old). Comparisons between groups were performed by using univariate analysis of variance (quantitative variables) and 2 or Fisher exact tests (categoric variables) depending on expected frequency. Differences with descriptive P values of .05 or less were considered statistically significant. Descriptive P values should not be used to make inferences in larger populations. All statistics were calculated by using software (SPSS for Windows, version 12.0.1; SPSS, Chicago, Ill).
Because the sets of images were not available for review, we made no attempt to confirm the accuracy of original interpretations. In every case, scans were read once by one of two radiologists; it was not possible to assess interobserver or intraobserver variability. We did not analyze potential interactions between different imaging findings.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
As expected, the proportion of patients with positive findings progressively increased with age (43% for patients younger than 40 years, 99% for patients 70 years or older) (Fig 1a). A similar pattern was seen for men (Fig 1b) and women (Fig 1c). A total of 3361 findings were reported: A mean of 2.8 findings per patient ± 2.1 (range, 1–12; median, 2) was observed. The mean number of findings per patient was similar (P = .278) for men (2.9 ± 2.1; range 1–12; median, 2) and for women (2.7 ± 2.1; range 1–12; median, 2).
|
|
|
|
|
|
|
|
A total of 779 thoracic findings were reported: a mean of 1.5 findings per patient ± 0.8 (range, 1–5; median, 1). Despite the difference in prevalence of findings, the mean numbers of findings in men (1.5 ± 0.7) and in women (1.7 ± 0.9) were similar (P = .9). Fifty-nine percent (461 of 779) of thoracic findings were in the lungs; 35% (271 of 779), in the mediastinum; and 6% (47 of 779), in the chest wall.
In total, 461 lung findings, which represented 59% of thoracic findings, were observed. The most common lung findings were parenchymal scars (153 [33%] of 461), indeterminate lung nodules (115 [25%] of 461), granulomas (71 [15%] of 461), and emphysema (42 [9%] of 461) (Table 4). Among patients with nodules, 70% (81 of 115) had a single nodule and 30% (34 of 115) had multiple nodules (exact number was not reported). In only 34 (30%) of 115 cases, nodule size was reported, and the mean size in these cases was 8 mm ± 4 (range, 2–25 mm). Sixty-three percent (45 of 71) of granulomas were single; and 37% (26 of 71), multiple (size and number were not reported). One patient had a 4-cm mass.
|
In total, 47 chest wall findings, which represented 6% of thoracic findings, were observed. Forty-three (91%) patients had breast findings, and four (9%) had chest wall lipomas. Most (41 [95%] of 43) breast findings were in women. These included asymmetric breast tissue (26 [63%] of 41), soft-tissue lesions (six [15%] of 41), nonspecific low-attenuating lesions (four [10%] of 41), and parenchymal calcifications (three [7%] of 41). Two (5%) of 43 breast abnormalities (both gynecomastia) were observed in men.
Abdominopelvic findings.—In total, 1517 abdominopelvic findings were observed. Sixty-nine percent (826 of 1192) of patients had at least one finding in the abdomen or pelvis. The proportion of patients with positive findings was similar (P = .66) for men (544 [70%] of 774) and for women (282 [67%] of 418). A total of 1517 abdominal findings were reported: A mean of 1.8 findings per patient ± 1 (range, 1–8 findings; median, 2) was observed. The mean number of findings in men (1.9 ± 1.0) was higher (P < .001) than that in women (1.6 ± 0.9). Three hundred ninety-nine patients had one abdominal finding; 253 patients, two findings; 115 patients, three findings; 38 patients, four findings; 14 patients, five findings; and seven patients, more than five findings.
The most common abdominopelvic findings involved blood vessels (561 [37%] of 1517), the genitourinary tract (364 [24%] of 1517), and the hepatobiliary tract (230 [15%] of 1517).
In total, 561 vascular findings, which represented 37% of abdominopelvic findings, were observed. By far, the most common vascular abnormalities were vessel wall calcifications (540 [96%] of 561). Aneurysms accounted for 2% (13 of 561); and vascular ectasia, for 1% (eight of 561). There were five aortic artery aneurysms (mean size, 32 mm; range, 15–48 mm), four iliac artery aneurysms (mean size, 19 mm; range, 13–23 mm), two renal artery aneurysms (size was not reported), and two splenic artery aneurysms (13 and 23 mm).
In total, 353 renal findings, which represented 23% of abdominopelvic findings, were observed. Renal findings are shown in Table 5. Fifty-six percent (58 of 104) of cysts were single; and 44% (46 of 104), multiple. Size was reported for only 14% (16 of 104) of cysts; the mean size was 37 mm ± 27, and the range was 10–100 mm. Sizes of indeterminate lesions were not reported.
|
|
In total, 84 female reproductive system findings, which represented 6% of abdominopelvic findings, were observed. Female reproductive system findings included uterine fibroids (36 [43%] of 84; measurements were not provided); ovarian cystic masses (27 [32%] of 84; mean diameter, 35 mm ± 13; range, 20–70 mm), with 23 considered physiologic, one considered possibly neoplastic, and three considered indeterminate; asymmetric adnexal enlargement (10 [12%] of 84; measurements were not provided); cervical cysts (seven [8%] of 84); and an enlarged or prominent cervix (four [5%] of 84; measurements were not provided).
In total, 62 splenic findings, which represented 4% of abdominopelvic findings, were observed. Eighty-seven percent (54 of 62) of splenic findings were calcifications, 11% (seven of 62) were splenic cysts or low-attenuating lesions (measurements were not reported), and 2% (one of 62) were mild splenomegaly (spleen dimensions were not reported). No solid or indeterminate splenic masses were described.
In total, 42 findings in the gastrointestinal tract and the mesentery, which represented 3% of abdominopelvic findings, were observed. Reported abnormalities included diverticuli (27 [64%] of 42), with 25 in the colon, one in the stomach, and one in the duodenum; nonspecific mural thickening (eight [19%] of 42), with seven in the colon and one in the stomach (measurements were not reported); appendicoliths without evidence of appendicitis (three [7%] of 42); and mesenteric calcifications (two [5%] of 42). In one patient (one [2%] of 42), an 8-cm mesenteric soft-tissue mass was observed, and in another, nonspecific mesenteric haziness without a soft-tissue component was observed.
In total, 30 pancreatic findings, which represented 2% of abdominopelvic findings, were observed. Pancreatic findings included fatty infiltration (16 [54%] of 30; mean age, 60 years; range, 42–76 years); diffuse pancreatic enlargement (seven [23%] of 30), with the possible cause and clinical importance not reported; soft-tissue mass suspicious for neoplasm (three [10%] of 30; mean size, 31 mm ± 7); and parenchymal calcifications (two [7%] of 30), which were possibly related to chronic pancreatitis. One patient had pancreatic duct dilatation (one [3%] of 30), with the diameter not reported and no pancreatic mass described; one patient had a pancreatic cyst (one [3%] of 30), with the size not reported.
In total, 26 abdominal lymph node findings, which represented 2% of abdominopelvic findings, were observed. Lymph node findings included calcified lymph nodes with normal size (11 [42%] of 26), overly numerous normal-appearing lymph nodes (eight [31%] of 26), and enlarged nodes (seven [27%] of 26). Of the seven patients with enlarged nodes, actual size was reported for only three (mean size, 16 mm ± 7; range, 11–25 mm); in one patient, the lymph nodes were described as mildly enlarged; in three patients, no measurements or subjective qualifiers were provided. The extent of enlargement was reported as diffuse in two cases and was not mentioned in five. One of the cases in which the extent of enlargement was reported as diffuse was considered suspicious for lymphoma. The other six cases had no reported differential diagnosis.
In total, 15 adrenal gland findings, which represented 1% of abdominopelvic findings, were observed. Fifty-three percent (eight of 15) of adrenal gland findings were adrenal gland calcifications. The other 47% (seven of 15) were focal lesions: Two were reported as adenomas (14 and 25 mm), and five were reported as indeterminate low-attenuating lesions (mean, 21 mm ± 6; range, 14–30 mm). Attenuation measurements were not reported.
Findings of the spine.—In total, 1065 findings of the spine were observed. Sixty percent (719 of 1192) of patients had at least one finding of the spine. Four hundred thirty-four patients had one finding of the spine; 226 patients, two findings; 57 patients, three findings; and two patients, four findings. The proportions of patients with positive findings were similar for men (477 [61%] of 774) and for women (242 [58%] of 418) (P = .57). A total of 1065 findings of the spinal were reported: A mean of 1.5 findings per patient ± 0.6 (range, 1–4 findings; median, 1) was observed. The mean numbers of findings in men (1.5 ± 0.6) and in women (1.4 ± 0.6) were similar (P = .9).
Findings of the spine included degenerative changes (583 [55%] of 1065); abnormally low bone density (416 [39%] of 1065); and miscellaneous benign lesions (63 [6%] of 1065), most commonly pars defects (n = 14) and vertebral hemangiomas (n = 11)). Three (<1%) of 1065 indeterminate sclerotic lesions were described; one was an exophytic lesion in the L5 transverse process and the other two were a focal lesion in L2 and a lesion in the left ilium.
Recommendations
Prevalence of recommendations.—Thirty-seven percent (445) of 1192 patients received one or more reported follow-up recommendations; 61% (273 of 445) were men, with a mean age of 58 years (range, 38–81 years), and 39% (172 of 445) were women, with a mean age of 58 years (range, 30–85 years). The proportion of patients with recommendations for follow-up progressively increased with age (11% at younger than 40 years and 56% at 70 years or older). A similar pattern was seen for both men and women (Fig 1). The proportion of recommendations was higher for women (172 [41%] of 418) than it was for men (273 [35%] of 774), but the difference was not significant (P = .17).
Recommendations per patient.—A total of 582 recommendations of 1192 were reported, with a mean of 0.5 recommendation per patient ± 0.7. Four hundred seventy-three (81%) of the 582 recommendations were stated in compulsory terms; the other 109 (19%) were suggested if clinically "appropriate" or "indicated." Three hundred thirty-seven patients received one recommendation; 86 patients, two recommendations; 16 patients, three recommendations; five patients, four recommendations; and one patient, five recommendations. The number of recommendations per patient was higher (P = .007) for women (0.6 ± 0.8) than it was for men (0.4 ± 0.7). The number of recommendations per patient progressively increased with age (P < .001) from 0.1 ± 0.3 (40 years or younger) to 0.8 ± 0.3 (70 years or older). A similar pattern was observed for both men and women (Fig 2).
Recommendations per finding.—The mean number of recommendations per finding was 0.17 ± 0.30. The mean number was greater (P = .001) for women (0.20 ± 0.40) than it was for men (0.15 ± 0.36); no significant difference was seen when the different age groups were compared. Figure 2 plots the number of recommendations per finding as a function of sex and age group. Recommendations per finding were significantly higher in women than they were in men for all age groups (P = .001). The larger number of recommendations per finding in women was largely influenced by ovarian and breast findings, as the findings in the ovaries (0.8 ± 0.3) and the breasts (0.6 ± 0.4) were associated with the highest number of recommendations per finding for any organ. After excluding reproductive organs and breasts, recommendations per finding were similar for men and for women overall and for each age group separately (data not shown). Table 3 shows the number of recommendations and recommendations per finding for each anatomic location.
Specific follow-up recommendations.—Follow-up imaging was the most common recommendation (400 [69%] of 582). Other recommendations are summarized in Table 7. Recommended imaging studies included CT (268 [67%] of 400), ultrasonography (US) (58 [14%] of 400), magnetic resonance (MR) imaging (42 [10%] of 400), nonspecified imaging follow-up (24 [6%] of 400), mammography (six [2%] of 400), upper gastrointestinal tract series (two [1%] of 400), and an option between CT and another examination (30 [8%] of 400). A specific imaging follow-up interval was given in only 42% (166 of 400) of cases. In those cases, the mean recommended interval was 8 months ± 5 (range, 1–60 months): 7 months ± 4 (range, 2–24 months) for thoracic findings and 10 months ± 6 (range, 1–60 months) for abdominal findings. Specific time intervals were not mentioned for spinal imaging recommendations. Imaging follow-up recommendations for abdominal findings are summarized in Figure 4.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
In our study, we described the frequency and spectrum of radiologic findings and follow-up recommendations in a population in which screening was performed and in which patients underwent whole-body CT screening at a for-profit community-based imaging center. Although physicians referred some patients, most (76%) were self referred. Unfortunately, the patient's motivation to undergo whole-body CT screening, despite its high cost, was not recorded by the imaging center. Although the study sample consisted mainly of middle-aged adults, there was a wide age range (22–85 years) among subjects who underwent screening. Four percent of patients who underwent screening as a result of self referral reported a history of malignancy (eg, colon, prostate, breast), medical disease, organ transplantation, or thoracoabdominal surgery. Since patients with known cancer are at higher risk for recurrent or metastatic disease, tailored diagnostic CT rather than whole-body CT screening is the preferred examination. Moreover, 8% were younger than 40 years old (including one who was 22 years old), and those in this age group have a low pretest probability of disease and, therefore, are unlikely to benefit from screening. Interestingly, a majority of patients who underwent screening were men. Male predominance was even higher in the self-referred subset. Further study is required to confirm the greater tendency of men to undergo screening and to delineate causative factors.
Eighty-six percent of patients who underwent screening had one or more findings, with exclusion of cardiac findings, at whole-body CT screening. The proportion of patients with findings and recommendations and the number of findings and recommendations per patient increased with age. Thus, findings and recommendations in patients younger than 40 years were unusual. On the other hand, prevalence of findings and follow-up recommendations were almost universal in patients older than 70 years. The clinical effect, positive or negative, of the findings and recommendations is unknown, because patient outcome has not yet been ascertained.
Investigators in prior studies (24–27) have reported the prevalence of findings in the chest and the abdomen at CT screening for carcinoma of the lung and the colon. In the chest, reported prevalence was 69% for indeterminate pulmonary nodules (3), 3% for emphysema (28), and 62% for other incidental abnormalities (29). In the abdomen, prevalence of findings with major clinical importance was 23% (26), while prevalence of any imaging finding was 87% (25). In another study in which findings in the abdomen and the chest were combined (3), the researchers reported that 79% of participants had at least one finding, usually a false-positive finding, that required further diagnostic testing. Comparison of our data with the data in these reported studies may not be valid, however, because of differences in patient selection and pretest probability of disease and lack of standardized reporting of findings.
Although a higher number of findings per patient was found in men, women had a significantly higher prevalence of follow-up recommendations and a higher number of recommendations per finding. These differences were predominantly driven by recommendations on the basis of findings in the breast and ovary. Overall, 37% of patients received follow-up recommendations; additional imaging (usually with CT) and consultation with a physician were recommended most frequently.
Downstream examinations generated from whole-body CT screening could have important financial ramifications for the individual and/or the health care system. Findings encountered at whole-body CT screening may conceivably alter patients' abilities to obtain medical, disability, or life insurance. Other adverse effects associated with work-up after positive results are obtained in asymptomatic patients who undergo screening include increased patient anxiety and potential complications from intravenous contrast material administration (30–32) or invasive procedures (eg, biopsy, surgery, endoscopy) that are performed for further characterization of encountered findings. These adverse consequences are particularly undesirable if the original screening results were false-positive or if the natural course of disease is not improved with early detection.
False-negative findings at whole-body CT screening may also have adverse consequences. On the basis of our anecdotal experience, we speculate that important disease may be missed at a whole-body CT examination without intravenous contrast material for which results are negative and could result in a false sense of security; these factors could potentially delay detection and treatment of actual disease. This may be particularly likely to occur in patients with a high risk for malignancy, such as those with a prior neoplasm or a family history of malignancy, or in symptomatic patients who pursue whole-body CT screening rather than physician-directed work-up. Because important abdominal abnormalities may be missed at unenhanced CT, use of whole-body CT screening with intravenous contrast material has been advocated (18). Whether the benefit of higher sensitivity and specificity outweighs potential complications from intravenous contrast material in asymptomatic patients, however, is unknown (19,20,33).
Another potentially adverse consequence of whole-body CT screening is radiation exposure. Although controversy exists as to whether radiation exposure increases the risk of malignancy (34), most authorities emphasize the importance of reducing patient exposure during radiologic examinations (21,22,35). Cumulative radiation exposure is especially of concern when young individuals undergo screening: They may require follow-up interval scanning on the basis of positive findings, or they may seek to obtain repeat scans on their own despite negative findings (34,36,37). Moreover, population exposure can be substantial in a screening population (particularly in young patients), especially when multiple follow-up studies cause additional radiation burden and thus increase the chance of cancer induction (21,22,38). Furthermore, it has been recently demonstrated that estimated radiation doses from single and repeated whole-body CT screening examinations correspond to a dose for which there is direct evidence of increased cancer mortality in atomic bomb survivors (38).
Unfortunately, unnecessary radiation exposure in young patients is likely with whole-body CT screening. Unlike chest CT screening for lung carcinoma and CT colonography for colon cancer, which are performed in high-risk individuals who are usually older than 50 years, whole-body CT screening is performed in unselected patients who may be young or otherwise at low risk of disease. Moreover, although low-dose protocols (reduced voltage and amperage parameters) are possible for chest and colon CT screening, low-dose protocols may be ineffective for whole-body CT screening because soft-tissue contrast is required for lesion detection in solid organs (20,23). Thus, CT lung screening (39) and CT colonography (40–42) both result in an estimated CT dose index of approximately 5 mGy/100 mAs, about one half the estimated dose index of electron-beam CT screening of the abdomen (10 mGy/100 mAs) (43). Thus, whole-body CT screening exposes young asymptomatic subjects with low disease risk to a higher radiation dose than does targeted screening in high-risk older individuals (23).
Recommended follow-up was inconsistent. For instance, recommended interval follow-up for lung nodules varied from 1 to 12 months, and no relationship between nodule size and interval was found (data not shown). Recommended imaging work-up for cystic ovarian lesions was also highly variable and included US, contrast material–enhanced CT, and MR imaging. These inconsistencies raise questions about effective communication, especially to self-referred patients, who comprised the majority of patients who underwent screening in our study. A standardized language of reporting findings and recommendations, similar to that used for mammography, would be useful. Moreover, there are many unanswered questions in regard to which lesions require follow-up (such as indeterminate lung nodules, asymmetric breast abnormalities, and subcentimeter low-attenuating lesions in solid organs), and long-term outcome studies are needed to standardize these criteria. Several issues must be addressed before whole-body CT screening is advocated for the general population. A high prevalence of detectable preclinical-phase disease is required for an effective screening program, otherwise a high number of false-positive results will be obtained (17,44–46). Even without knowing the final outcome of the patients included in our study, the low prevalence of findings and recommendations in younger patients suggests that whole-body CT screening is unlikely to be beneficial in this age group. Moreover, the extremely high prevalence of findings and the need for additional work-up in older patients also limits the benefit of screening the elderly, as virtually all older patients require additional work-up with more definitive tests. Thus, whole-body CT screening is likely to be beneficial, if at all, only in patients in well-defined age or risk categories.
Moreover, for screening to be cost effective, two, among other, criteria must be met: (a) screening must help in the detection of preclinical disease and (b) detection of preclinical disease must translate into longer survival and/or higher quality of life. Several legal, social, and health care issues in regard to whole-body CT screening must also be addressed. These include reaction to intravenous contrast material in cases in which it is used for screening (19); the legal impact of false-negative examinations (23,47), especially when potentially curable life-threatening disease is missed; complications after follow-up examinations or interventions for further characterization of false-positive whole-body CT screening findings (10); lack of compliance with follow-up recommendations (eg, because the report was never received by the intended recipient or because the patient did not understand the instructions); radiation exposure; impact on health, disability, and life insurance; and who should pay for further work-up as a result of whole-body CT screening findings in self-referred patients.
A limitation of our study was that it was a retrospective review of radiologic reports. Also, we were unable to confirm the reported findings by reviewing the images themselves. A panel was necessary to codify inconsistently worded findings and recommendations. Interobserver and intraobserver variability could not be analyzed. Although our purpose was to describe imaging findings and recommendations after whole-body CT screening, lack of final diagnosis and outcomes for detected abnormalities constitutes a major limitation. As a result, the test performance for whole-body CT screening cannot be established at this time. Even if final clinical outcomes were known, the absence of a control group (patients who did not undergo screening) prevents assessment of mortality reduction and the effect on the quality of life. In theory, whole-body CT screening offers an exciting opportunity to establish the prevalence of findings in asymptomatic patients, although our results are difficult to generalize to the entire U.S. population or even to the population of southern California, where the screening was performed. The sampled patients were not chosen at random. Instead, patients sought and paid for the examinations and, therefore, independent of their motivation, probably represent a population with a higher socioeconomic status than that of the general population. Moreover, patients' ethnic backgrounds were unknown.
On the basis of our results, we believe that, prior to advocating whole-body CT screening to the general public, further research is needed to optimize whole-body CT screening techniques, determine whole-body CT screening diagnostic performance, standardize reporting, establish a system of patient stratification according to risk and demographic factors, fully address ethical and legal issues (48), create cost-benefit models, and, if still warranted, perform randomized clinical trials.
We conclude that there is a high prevalence (86%) of reported findings with whole-body CT screening and that most are benign by description; however, 37% of patients receive recommendations for follow-up examinations, most frequently for findings in the lungs and the kidneys. Thus, we recommend that patients who undergo whole-body CT screening should be informed prior to the procedure that additional tests may be necessary in up to 37% of cases, since this possibility can increase interval anxiety rather than provide the reassurance they are seeking. Moreover, the probability of findings and the need for additional follow-up increases significantly with age, so patients in the oldest age groups are almost guaranteed to require ancillary tests. Patients should also be counseled that the negative predictive value of whole-body CT screening is unknown and that several complex issues have not yet been addressed. Whether the detection of a wide spectrum of findings with whole-body CT screening can provide patients with real benefits from early diagnosis of disease or, if scans are negative, valid reassurance remains uncertain.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, C.D.F., D.A.A., G.C.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, C.D.F., D.A.A., D.D., S.K.S., P.L., F.S., G.C.; clinical studies, D.A.A., D.D., S.K.S., P.L., F.S., G.C.; statistical analysis, D.A.A., C.B.S., D.D., D.L.L., G.C.; manuscript editing, C.D.F., D.A.A., C.B.S., M.A.B., D.L.L., G.C.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. J. Pickhardt, M. E. Hanson, D. J. Vanness, J. Y. Lo, D. H. Kim, A. J. Taylor, T. C. Winter, and J. L. Hinshaw Unsuspected Extracolonic Findings at Screening CT Colonography: Clinical and Economic Impact Radiology, October 1, 2008; 249(1): 151 - 159. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. M. Burger, N. E. Kass, J. H. Sunshine, and S. S. Siegelman The Use of CT for Screening: A National Survey of Radiologists' Activities and Attitudes Radiology, July 1, 2008; 248(1): 160 - 168. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E J HALL and D J BRENNER Cancer risks from diagnostic radiology Br. J. Radiol., May 1, 2008; 81(965): 362 - 378. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Schoder and M. Gonen Screening for Cancer with PET and PET/CT: Potential and Limitations J. Nucl. Med., January 1, 2007; 48(1_suppl): 4S - 18S. [Abstract] [Full Text] [PDF]
|
|
) and of patients with follow-up recommendations (
) over the total number of patients per age group. (a) Overall, the proportion of patients with positive findings and recommendations progressively increased with age, and a similar pattern was noted in (b) men and (c) women. Proportions of recommendations were higher for women than for men, although the difference was not significant (P = .19).
) and per finding (
) over the total number of patients per age group. (a) Overall, the number of findings and recommendations per patient increased as the age increased; the number of recommendations per finding was similar for all age groups, except for a nonsignificant increment in middle-aged patients (40–49 and 50–59 years old) (P = .09). Similar patterns were seen for (b) men and (c) women. Women had a higher number of recommendations per finding (P = .001) and per patient (P = .14) than did men in all age groups.
