| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
University of Illinois College of Medicine (A.B.S., B.C.) and Columbia-Michael Reese Hospital (A.B.S., C.B., J.L., J.R., H.H., E.S.-F., T.G.), Chicago, Illinois 60612
Address all correspondence and requests for reprints to: Arthur B. Schneider, M.D., Ph.D., Section of Endocrinology and Metabolism (MC 640), University of Illinois College of Medicine, 1819 West Polk Street, Chicago, Illinois 60612.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
1.0-cm nodules) were
the only nodules of this size in 7 subjects. Of 11 nodules 1.5 cm or
larger, only 5 were palpable. Serum thyroglobulin correlated to the
number (P = 0.04; r2 = 0.10), but not
the volume of the thyroid nodules (P = 0.07;
r2 = 0.08). We conclude that thyroid nodules are continuing
to occur and are exceedingly common in this irradiated cohort of
individuals. The results confirm that thyroid ultrasonography is more
sensitive than physical examination and scanning. However, thyroid
ultrasound is so sensitive and nodules so prevalent that great caution
is needed in interpreting the results. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
As thyroid ultrasound can detect nodules as small as 2–3 mm (13, 14) and because no radiation is involved, it has gained wide acceptance for the screening of nodular thyroid disease. The previous reluctance to use scintigraphy has been replaced with the frequent use of thyroid ultrasound, e.g. in the Chernobyl region (15, 16). This raises the concern that its extremely sensitive findings will be misused in planning the further evaluation, treatment, and follow-up of patients (17, 18).
The purpose of the current study was to study the role of thyroid ultrasonography. Since 1974 we have been following a cohort of 4296 individuals who received childhood radiation treatment for benign conditions during the period between 1939–1962 (6, 19, 20). By January 1, 1997, 355 thyroid cancers had been diagnosed in the 3058 subjects who had been located. We invited subjects from a previously identified subgroup who had been seen between 1974–1976 and who had normal thyroid physical examinations and thyroid scans at that time (20). Those living in the Chicago area, without thyroid surgery in the interval, and with accurate thyroid-specific radiation doses were considered eligible. Thyroid ultrasounds, 99mTc-pertechnetate scans, and palpation were performed on 54 participants. We describe the continuing occurrence and the striking frequency of nodular thyroid disease in irradiated individuals.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Subjects were members of a cohort of patients who were treated with radiation for benign conditions in the head and neck area between 1939 and 1962. Between 1974–1976, 1476 individuals were screened by palpation of the thyroid, a scan using 99mTc-pertechnetate with a pinhole collimator and measurement of serum thyroglobulin (Tg) (19). We identified 115 patients with normal exams, normal scans, and elevated serum Tg levels (high Tg group). A second group of the same size came from the remaining 796 normal patients with normal serum Tg levels (normal Tg group). The 115 were selected from the 796 using a table of random numbers. These are the same groups that were the basis for a prior study evaluating screening with isotopic scans (11).
To define eligible subjects, these 2 groups were diminished, as follows (numbers in parentheses are reductions for the high Tg and normal Tg groups, respectively): originally misclassified (4, 2), not part of the subsequently redefined cohort that is limited to conventional radiation treatment before the 16th birthday (12, 17), thyroid surgery since original screening (17, 14), dead (5, 3), and last known address not in the Chicago area (15, 16). The resulting 62 subjects in the high Tg group and the 63 in the normal Tg group were solicited by letter and follow-up telephone calls.
Of the 62 eligible in the high Tg group, 22 (35.5%) participated in this study, and of the 63 eligible in the normal Tg group, 32 (50.8%) participated. The reasons for nonparticipation were as follows (numbers in parentheses are nonparticipants in the high Tg and normal Tg groups, respectively): lost to follow-up (3, 4), no response (18, 14), out of area (3, 0), deceased (3, 0), and refused, deferred, or partial participation (14, 13).
Clinical evaluation
The evaluations included a health survey, palpation of the thyroid and salivary glands, a 99mTc-pertechnetate thyroid scan as previously described, and a thyroid ultrasound examination (11, 19). The study was reviewed and approved by the local Institutional Review Board, and written informed consent was obtained from each of the participants. The ultrasound was performed with a 7.5-MHz transducer in direct contact with the lubricated skin of the extended neck. A minimum of six transverse and six longitudinal static images were obtained for each lobe by a registered diagnostic medical sonographer. A board-certified radiologist reviewed the examination and obtained additional images as necessary. Tg levels were determined using a RIA with an upper limit of normal of 28 ng/mL (21, 22). Briefly, an antibody to human Tg was raised in rabbits, and a double antibody precipitation method was used. The precipitating antibody was unreactive with human Igs, avoiding falsely decreased values in the presence of endogenous anti-Tg antibodies. The assay was calibrated for the level of anti-Tg antibodies, measured by immunoprecipitation, that interferes with the results. Four samples with interfering anti-Tg antibodies were excluded.
Ultrasound and scan reviews
Ultrasound. Each ultrasound exam was read independently by
two radiologists (J.L. and J.R.) and one endocrinologist (A.S.). Each
nodule was characterized as follows: 1) presence; certain or uncertain
(uncertain refers to an abnormal area that was not clearly demarcated
from surrounding thyroid tissue); 2) size in three planes (in
millimeters); 3) location; anterior or posterior; 4) location; upper
pole, midportion of lobe, lower pole, or isthmus; and 5) type; solid,
cystic, or mixed. A consensus was arrived at differently for large and
small nodules. For nodules described as 10 mm or more in largest
dimension by one or more readers, a nodule by nodule consensus was
reached. For each nodule, each reader’s interpretation was scored as
certain (C), uncertain (U), or absent (A). The consensus was recorded
as certain for CCC and CCU. The consensus was recorded as uncertain for
CUU, UUU, and UUA. The consensus was recorded as absent for UAA and
AAA. If the initial readings were CCA, CUA, and CAA (at least one
certain and one absent), then the three readers reached a consensus as
a group. For small (<10 mm) nodules, each reader was classified as
having found zero, one, or two or more of them. If the three readers
were not in agreement, then the ultrasound examination was reviewed as
a group. The volume of each nodule was estimated as length x
width x breadth x 
/6.
Scan. Thyroid scans were read independently by two nuclear medicine specialists and one endocrinologist (C.B., H.H., and A.S., respectively). First, each reader reviewed each subject’s most recent scan. Then they reviewed all of the subject’s scans, including those performed between the first and the current scan, and recorded 1) an interpretation of each scan, 2) progression of abnormal areas between scans, and 3) progression of abnormal areas from the first to the last scan. Abnormal areas were characterized as follows: 1) presence; certain or uncertain (uncertain refers to an abnormal area that was not characteristic of a thyroid nodule); 2) size; small, medium, or large (corresponding to thirds of a thyroid lobe); 3) location; upper lobe, midportion of lobe, lower lobe, or isthmus; 4) type; cold, warm, or hot; and 5) progression between first and last scans on a scale of -3 to +3, where zero is no progression and 1, 2, and 3 are minimal (or uncertain), definite, and marked progression. Negative values for progression (i.e. regression) were permitted to allow for the possibility that an original scan would, retrospectively, appear abnormal. The consensus for each abnormal area on the scan was reached using the same procedure as for 10 mm or larger ultrasound-detected nodules. For abnormal areas on the first and last scans, each reader’s interpretation was scored as certain (C), uncertain (U), or absent (A). The consensus progression score was taken as the average of the three readers when they were in agreement and was assigned by a group reading when they were not (when at least one reader found definite or marked progression and at least one reader found no progression).
Statistical tests
The correlation of the ultrasound and scan findings was based on
the final consensus readings for each. Group comparisons were made
using 
2 analysis or Student’s t test. The
kappa value was used for measuring the concordance of ultrasound and
scan readings (23). This statistic takes into account the concordance
in readings that would occur even if the readings were completely
random. It is equal to 1.0 for complete concordance and to zero for
random readings. These tests and correlation coefficients were
calculated using NCSS 6.0 (NCSS Statistical Software, Kaysville, UT)
software.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
There were no significant differences (by t test or
2 testing) for any of the factors comparing the
participants to the nonparticipants or to the entire cohort (Table 1
). Table 1 includes all of the factors
that are known to be independent risk factors for developing benign and
malignant thyroid nodules (6, 24).
|
Thyroid ultrasound examinations
Only 7 of the 54 subjects had normal thyroid ultrasound
examinations (Table 2). A total of 157
certain thyroid nodules were detected by the ultrasound examinations.
Forty of them were 1.0 cm or greater in largest dimension and were
found in 28 subjects. The remaining 117 nodules were less than 1.0 cm
in largest dimension and occurred in 45 subjects. The size distribution
(in 0.1-cm increments) of the ultrasound-detected nodules was nearly
uniform up to 0.8 cm and then declined gradually (shown by the total
heights of the bars in Fig. 1). The
largest nodule measured 3.0 cm in the greatest dimension.
|
|
). The largest
number, 6, was present in 5 subjects. To determine whether there was a
relationship between large and small nodules, the frequency of small
nodules was compared on the basis of whether large nodules were also
present (Fig. 2). The presence and frequency of small nodules were not
related to the presence of large nodules.
|
Three of the original 54 scans were interpreted as having single abnormal areas with decreased uptake, whereas no abnormalities were found in the other 51. The 3 changes in interpretation were due in part to the availability of the more recent scans.
For the recent scans, there were 46 certain abnormal areas in 29
subjects (Fig. 3). Two of these were
warm, and 1 was hot. For these 29 subjects, the progression scores were
minimal or uncertain (1.0–1.3) in 6, definite (2.0–2.3) in 16, and
marked (2.7–3.0) in 7. In addition, on the recent scans there were 20
areas in 12 subjects that were considered possible, but not definite,
nodules. One of the 3 original scans that were not normal regressed to
normal. Progression or regression could not be judged for ultrasound
examinations as previous scans had not been performed.
|
The larger the ultrasound-detected nodule, the more likely that it
was present on the scan (Fig. 1
, black and cross-hatched
areas). Some of the smaller ultrasound-detected nodules were near
larger nodules or occurred in groups, allowing them to be correlated to
scan abnormalities.
The thyroid ultrasound examinations showed both large and small nodules
that were not present on the scans (Table 3). Of the 40 certain ultrasound-detected
1.0-cm or larger nodules, 10 (25%) were not present on the
corresponding scans. Of the 117 ultrasound-detected small nodules, 91
(77.8%) were not present on the corresponding scan.
|
). In the
remaining 6 cases, the scan abnormalities were all in the smallest
category, and 1 of the patients had been transiently anti-Tg positive.
A large ultrasound-detected nodule, matching another scan abnormality,
was present in the same lobe in 2 of these cases. In another case there
were multiple, small, ultrasound-detected nodules in the same lobe, but
none clearly matched. In 3 cases the corresponding lobes showed no
nodules, but there were varying degrees of diffuse heterogeneity. For
uncertain scan areas (Table 3), 7 (38.9%) of 18 cases were associated
with ultrasound-detected nodules. The overlap of ultrasound-detected nodules 1.0 cm or larger and abnormal scans in individuals was as follows. Of the 28 individuals with ultrasound-detected 1.0-cm or larger nodules, 20 also had abnormal scans, 2 had scans with uncertain abnormalities, and 6 had normal scans. Of the 29 individuals with abnormal scans, 20 also had ultrasound-detected 1.0-cm or larger nodules, 7 had ultrasound-detected less than 1.0-cm nodules (largest dimensions of 9, 9, 8, 8, 4, 3, and 2 mm, respectively), and 2 had normal ultrasound examinations.
Reproducibility of readings
The kappa statistic was used to quantify the reproducibility of
the image readings between pairs of readers and between each reader and
the final consensus (Table 4). The
agreement for the number of large (1.0 cm) ultrasound-detected
nodules was very good, with a range of 0.66 ± 0.10 to 0.88
± 0.10 for the each of the readers compared to the consensus. The
agreement was equally good for the number of small nodules, with the
corresponding range of 0.63 ± 0.09 to 0.88 ± 0.10. For the
thyroid scans divided into normal scans and scans that were abnormal or
uncertain, the reproducibility was comparable to that for ultrasound
with a range of 0.47 ± 0.12 to 0.79 ± 0.13 for the
comparison to the consensus. When three categories were used to
classify the scans, the range of kappa values expanded to 0.29 ±
0.09 to 0.80 ± 0.10.
|
Although it is generally held that most thyroid nodules 1.5 cm or
larger are palpable, of the 11 subjects with ultrasound-detected
nodules 1.5 cm or larger, only 5 were palpable. The palpable ones (Fig. 4, left) were anterior in
location and protruded from the gland. The nonpalpable ones (Fig. 4, right) were more posterior and caused less distortion of the
gland contour. All but 1 of the 1.5-cm or larger nodules were seen on
scans.
|
About 6 months after their evaluation we requested follow-up
information. Of the 11 subjects with ultrasound-detected nodules 1.5 cm
or larger, histological data were available for 6 of them (Table 5). None of the 5 fine needle aspirations
were suggestive of malignancy, although 2 mentioned the possibility of
a neoplasm, and 1 noted atypical nuclear features. In 1 instance, a
subject died of unrelated causes, and an autopsy showed that the large
nodule was a papillary cancer.
|
We previously reported on the initial follow-up of the high Tg and
normal Tg groups (11, 25). At 9 yr of follow-up, a significantly higher
frequency of thyroid nodules had occurred in the high Tg group. The
extension of this follow-up to the time of the present study (but not
including the results of this study) shows that the difference was not
sustained beyond about 15 yr (Fig. 5).
|
). Correlation was made to the number of
nodules per subject and to the total volume of nodules. Neither of
these correlated with the original Tg levels. The study Tg levels
correlated with the number of nodules (P = 0.04;
r2 = 0.10; intercept = 2.2 nodules; slope =
0.012), but not with the total volume (P = 0.07;
r2 = 0.08; intercept = 0.43 cm3;
slope = 4.1).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This study shows how the advantages and problems of thyroid ultrasound are magnified in irradiated patients. The major advantage is that ultrasound can find clinically important, potentially malignant, nodules that are not otherwise detectable. Here, approximately half of the nodules 1.5 cm or larger were not palpated, and 25% (10 of 40) of the nodules 1.0 cm or larger were not detected by thyroid scanning. Similar palpation results have been observed by others (17, 26, 27, 28, 29). Ezzat et al. (17) found no palpable nodules in 16% of subjects with ultrasound-detected nodules larger than 1.0 cm; Tomimori et al. (26) could palpate the nodules in only 7 of 18 patients with nodules between 1.5–4.8 cm; Brander et al. (27) found 14 nonpalpable nodules larger than 2.0 cm in 72 subjects; Witterick et al. (28) found 15 nonpalpable nodules larger than 1.5 cm in 60 patients undergoing thyroid surgery; and Price et al. (29) found 17 nonpalpable nodules, some described as large cancers, in 55 patients undergoing thyroid surgery. Although the observed frequency of large, ultrasound-detected, palpation-negative nodules varied, it is clear that many large, ultrasound-detected nodules escape detection by palpation.
The results reported here can be compared to several reports on the prevalence of ultrasound-detected thyroid nodules in adult populations not exposed to radiation (17, 18, 26, 27, 30, 31). The range for the prevalence of nodules found in these reports is 13–67%. The upper end of this range comes from a study that has the unusual feature that 21% of the patients subjected to thyroid ultrasound screening had palpable thyroid nodules (17). The reported prevalence is higher in women and increases with age. The presence of a palpable nodule is associated with additional ultrasound-detected nodules in about half of the cases (32). The comparison indicates that the prevalence of ultrasound-detected nodules is increased in the irradiated population studied here.
The sensitivity of a diagnostic test such as ultrasonography gives rise to the possibility that unneeded additional tests and surgery may be performed (10). This concern is magnified in people with a history of radiation exposure. Nagataki et al. (33) screened 2587 individuals who were in Nagasaki at the time of the atomic bomb. Thyroid disease was found in 477 (17.3%), in 260 as a result of the screening, mostly by ultrasound. In 19 there were nodules that were large enough for fine needle aspiration biopsies. Antonelli et al. (34) screened 50 hospital workers with occupational exposure to radiation and 100 controls with ultrasound. Newly detected nodules were present in 19 (38%) exposed workers (10 with at least 1 nodule >1.0 cm) compared to 13 (13%) in unexposed workers (5 with at least 1 nodule >1.0 cm). Ito et al. (15) screened 55,054 Chernobyl area children with ultrasound and found that 1,396 were abnormal, and 197 had nodules larger than 5 mm. Sugenoya et al. (16) compared 299 children, aged 10–15 yr, from Chernobyl-exposed areas to 323 unexposed children. None had palpable thyroid abnormalities, but 34 exposed children and only 4 unexposed children had ultrasound-detected thyroid abnormalities. These studies confirm that small, clinically unapparent, ultrasound-detected nodules are more frequent in radiation-exposed individuals (6).
The second disadvantage of thyroid ultrasound is that it is an operator-dependent test. In a study in which 2 independent observers performed thyroid ultrasound examinations on 76 subjects (152 lobes), the kappa value for agreement was 0.55–0.60 (35). A solitary solid lesion was observed in 22 lobes, but the observations were concordant in only 10 (45%) of them. Here, the kappa value for large nodules was higher, but readings were performed on the recorded images, where a higher concordance would be expected. The kappa values for ultrasounds and scans were similar, but the number of uncertain findings in the scans was greater.
For 6 of 46 scan abnormalities, a matching nodule was not found by ultrasound. These areas with abnormal isotope uptake may be due to focal, nonneoplastic changes, such as thyroiditis. However, nascent nodule formation, especially in the cases where other nodules were present in the same lobe, cannot be excluded.
High Tg levels are associated with an increased chance of developing thyroid nodules (11, 25). The extended follow-up reported here shows that the excess risk associated with a high Tg level persists for about 10 yr. Tg levels that rise over time are also associated with developing nodules (36). In this study we saw a significant, but small, correlation of Tg levels to the number, but not the volume, of thyroid nodules.
How should the available methods for following patients at risk for developing thyroid neoplasms be applied? First, it should be decided whether a given individual’s risk factors, e.g. dose and age at treatment, are high enough to warrant imaging (1, 4). If so, the findings support the use of thyroid ultrasound at intervals determined by the clinical findings. We suggest every 3–5 yr if the findings are normal. For patients who have had a series of thyroid scans, an additional scan may be useful to judge whether any changes have occurred. A scan abnormality that is not seen by ultrasound suggests the need for more careful follow-up, but it is not clear whether follow-up scans, in addition to ultrasound, are helpful. For nodules 1.0–1.5 cm or larger, we recommend fine needle aspiration, with ultrasound guidance if necessary. For irradiated patients with benign thyroid nodules, thyroid hormone reduces the chance of recurrence (37). Its effectiveness in preventing nodules in the first place and preventing an increase in size of a preexisting one has been more difficult to show (1, 4). We favor thyroid hormone treatment for those at especially high risk, but this approach has not been widely adopted. To date there is no evidence that patients with nodules and childhood radiation should be treated differently from other patients. These issues need continued attention, as thyroid malignancies are more aggressive in older individuals.
|
Footnotes |
|---|
Received April 24, 1997.
Revised July 9, 1997.
Revised August 6, 1997.
Accepted August 12, 1997.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Imaizumi, T. Usa, T. Tominaga, M. Akahoshi, K. Ashizawa, S. Ichimaru, E. Nakashima, R. Ishii, E. Ejima, A. Hida, et al. Long-Term Prognosis of Thyroid Nodule Cases Compared with Nodule-Free Controls in Atomic Bomb Survivors J. Clin. Endocrinol. Metab., September 1, 2005; 90(9): 5009 - 5014. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Momani, E. Shore-Freedman, B. J. Collins, J. Lubin, E. Ron, and A. B. Schneider Familial Concordance of Thyroid and Other Head and Neck Tumors in an Irradiated Cohort: Analysis of Contributing Factors J. Clin. Endocrinol. Metab., May 1, 2004; 89(5): 2185 - 2191. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Bucci, E. Shore-Freedman, T. Gierlowski, D. Mihailescu, E. Ron, and A. B. Schneider Behavior of Small Thyroid Cancers Found by Screening Radiation-Exposed Individuals J. Clin. Endocrinol. Metab., August 1, 2001; 86(8): 3711 - 3716. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Marqusee, C. B. Benson, M. C. Frates, P. M. Doubilet, P. R. Larsen, E. S. Cibas, and S. J. Mandel Usefulness of Ultrasonography in the Management of Nodular Thyroid Disease Ann Intern Med, November 7, 2000; 133(9): 696 - 700. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
