Endocrine Sections, Washington Hospital Center and Georgetown University, Washington, D.C. 20010
Address all correspondence and requests for reprints to: Dr. Kenneth D. Burman, Chief, Endocrine Section, Washington Hospital Center, 110 Irving Street NW, Washington, D.C. 20010. E-mail: Kenneth.Burman{at}Medstar.net.
The contentious question regarding who should have calcitonin testing for the evaluation of early sporadic medullary thyroid carcinoma just got a little more confusing. Medullary thyroid carcinoma is a rare calcitonin-secreting neoplasm that occurs in both a familial and sporadic form. Seventy-five percent of medullary cancers are sporadic, whereas the remaining 25% occur as part of either the familial medullary thyroid cancer or multiple endocrine neoplasia type 2 syndromes (1). Medullary thyroid cancer represents approximately 4% of all thyroid malignancies. When medullary thyroid cancer presents with nodular disease, it is frequently metastatic, and the prognosis is usually poor. Early diagnosis and definitive treatment with total thyroidectomy, before the development of lymph node or distant metastases, offers the best hope of cure (2, 3, 4). It has been suggested that calcitonin testing, both with and without pentagastrin stimulation, may facilitate the early diagnosis of this disease and ultimately decrease morbidity and mortality (5, 6, 7, 8, 9).
In this issue of JCEM, Karanikas et al. (10) examine a series of 414 patients both with and without nodular thyroid disease and enter this ongoing debate. Of their 414 patients, 28 had abnormal basal calcitonin levels. Four of these patients also had elevated calcitonin levels greater than 100 pg/ml after pentagastrin stimulation. Of these four patients, only one went to surgery and had a confirmed diagnosis of medullary thyroid cancer.
Another interesting finding in their patient population is the incidence of abnormal calcitonin levels in the presence of Hashimotos thyroiditis. Although the prevalence of Hashimotos thyroiditis was 13.3% in the entire 414 patient cohort, among those with abnormally elevated serum calcitonin levels, the prevalence of Hashimotos thyroiditis was 25%. In contrast, 81 of 188 (43%) subjects with non-nodular thyroid disease in a series reported by Rieu et al. (8) had Hashimotos thyroiditis with normal basal calcitonin levels.
Karanikas et al. (10) correctly emphasize the paucity of data that relate non-nodular thyroid disease and calcitonin levels. This is the first substantive study that demonstrates a significant relationship between non-nodular thyroid disease, specifically Hashimotos thyroiditis, and abnormal calcitonin levels. They demonstrate not only a statistically comparable degree of calcitonin elevation in patients with and without nodular thyroid disease but also demonstrate that patients with Hashimotos thyroiditis are more likely to have abnormal calcitonin levels than others with non-nodular thyroid disease. Based on these observations, Karanikas et al. (10) have proposed extending the use of calcitonin testing to identify occult medullary cancer not only in patients with nodular thyroid disease but also in patients with Hashimotos thyroiditis. Is this recommendation prudent?
There is a probable correlation between Hashimotos thyroiditis, C-cell hyperplasia, and thyroid carcinoma (11, 12, 13). In the Rieu series, 11.3% of the patients with nodular thyroid disease had Hashimotos thyroiditis (8). Pacini et al. (9) also made specific mention of the inclusion of patients with Hashimotos thyroiditis among those with nodular thyroid disease. It is not clear how many patients from these two series were subsequently diagnosed with medullary thyroid cancer. Medullary thyroid carcinoma coincident with thyroiditis was found in half of the 16 patients with elevated basal calcitonin levels reported by Niccoli et al. (14).
Most of the data related to the use of calcitonin levels in the detection of early sporadic medullary carcinoma has been accumulated in patients with nodular thyroid disease. Approximately 0.51.5% of patients with nodular thyroid disease have medullary thyroid cancer (5, 6, 7, 8, 9, 14). Hahm et al. (5) studied 1448 patients and identified 56 patients (3.8%) with elevated serum calcitonin levels. Ten of these patients (0.7% of total) were proven to have medullary thyroid cancer. Of these 10 patients, basal calcitonin levels varied. However, all 10 patients had calcitonin levels after pentagastrin stimulation of greater than 100 pg/ml. Most importantly, of these 10 patients, only five had a fine-needle aspiration (FNA) that provided an indication for surgery. In the remaining five patients, the presence of an abnormal calcitonin level was the only indication for surgery and subsequent discovery of occult medullary cancer. Ozgen et al. (6) identified four medullary thyroid cancers in 773 patients with nodular thyroid disease (0.5%). Although FNA provided an indication for surgery in three of the four patients, none of the four FNAs showed medullary cancer. In this series, one patient with medullary cancer was identified solely because of an elevated calcitonin level. Vierhapper, Rieu, Pacini, and colleagues (7, 8, 9) reported similar results. Vierhapper et al. (7) found elevated basal calcitonin levels in 55 of 1062 patients (5.1%); 14 of these patients had pentagastrin-stimulated calcitonin levels greater than 100 pg/ml. Twelve of these 14 patients ultimately went to surgery. One patient underwent total thyroidectomy and was found to have only benign nodular disease. Eleven of the 12 (1% of total) that underwent surgery had either medullary thyroid cancer or C-cell hyperplasia. Eight patients in this series with elevated calcitonin levels underwent FNA. Again, in this series, FNA provided an indication for surgery in only four of these eight patients. FNA demonstrated medullary thyroid carcinoma in only two of these patients. The remaining patients were taken to surgery solely as a consequence of abnormal stimulated calcitonin levels. Rieu et al. (8) showed four of 469 patients with nodular thyroid disease (0.9%) had medullary thyroid cancer. One patient had histologically proven medullary thyroid cancer. Three others had C-cell hyperplasia. Each of these four patients not only had basal calcitonin levels that were elevated but also had pentagastrin-stimulated calcitonin levels greater than 100 pg/ml. Pacini et al. (9) noted eight of 1385 patients with nodular thyroid disease (0.6%) had elevated calcitonin levels and pentagastrin-stimulated calcitonin levels greater than 100 pg/ml. All were demonstrated to have medullary thyroid cancer. FNAs suggested medullary thyroid cancer in only two of these eight patients. In a retrospective series by Niccoli et al. (14), 34 of 1167 (2.9%) patients had elevated calcitonin levels. Fourteen of these 34 patients (1.2% of total) had medullary thyroid cancer. Preoperative FNA was diagnostic in only three of these cases.
As demonstrated by these studies, FNA suggests the diagnosis of medullary thyroid cancer less than half of the time. Many of these aspirations are, in fact, read as benign. Based on these studies, an elevated calcitonin level on routine screening of patients with nodular thyroid disease is much more likely to suggest medullary thyroid cancer than does an FNA. Yet, only 1040% of elevated basal calcitonin levels actually correspond to subsequently diagnosed medullary thyroid cancer.
False positive calcitonin elevations (i.e. elevation of calcitonin in the absence of medullary cancer) also occur. Benign C-cell hyperplasia, benign nodules, differentiated thyroid cancer, or less likely undiagnosed medullary thyroid cancer have been reported. Nonthyroidal causes have also been implicated and include sepsis or generalized inflammation, neuroendocrine tumors (e.g. lung, breast, prostate), and possibly exercise (15). Vierhapper et al. (16) recently described a series of 8374 patients with nodular thyroid disease that used calcitonin testing. Over 50% of 56 patients with basal or pentagastrin-stimulated calcitonin levels greater than 100 pg/ml who underwent total thyroidectomy were shown to have C-cell hyperplasia. Only 27 of these 56 patients were shown to have medullary thyroid cancer (16).
The largest body of literature examining the role of calcitonin testing in the diagnosis of medullary thyroid carcinoma exists in the several European studies cited previously (5, 6, 7, 8, 9, 14). Those studies, with the exception of Ozgen et al. (6), have used a strategy that uses a basal calcitonin level as an initial screen, followed by pentagastrin-stimulated calcitonin levels as a more specific confirmatory test. The use of basal calcitonin testing without follow-up pentagastrin stimulation as the sole means by which to detect medullary thyroid carcinoma is less clearly established. However, this is precisely the information of greatest interest to many in the United States because pentagastrin in no longer available in this country.
There is no absolute threshold value for basal calcitonin that discriminates benign from malignant disease. Furthermore, it is difficult to compare studies over time because they have historically used different assays of varying sensitivity. Ozgen et al. (6) found four patients in their series with medullary thyroid carcinoma without the use of pentagastrin stimulation. All patients in this study with medullary thyroid cancer had basal calcitonin levels above 100 pg/ml (range, 150-1000). Niccoli et al. (14) reported retrospectively that all patients in their series with a basal calcitonin level greater than 35 pg/ml had subsequently confirmed medullary thyroid cancer on postoperative pathology. Scheuba et al. (17) have shown that a basal calcitonin level of 64 pg/ml or a pentagastrin-stimulated calcitonin level of more than 560 pg/ml detected 31 of 38 patients (82%) with sporadic medullary thyroid carcinoma. Although it is extremely difficult to derive generalizations from the literature, Dr. Robert Gagel has noted that a threshold value for basal calcitonin of 25 pg/ml will detect approximately 90% of patients with medullary thyroid cancer. Although, as he notes, many false positive results occur within this range (Gagel, R., personal communication). We also agree with Dr. Gagels identification of basal calcitonin values between 10100 pg/ml as a gray zone. The predictive value of basal calcitonin levels in this range vary dramatically. Across all the series we have analyzed, basal calcitonin values between 10100 pg/ml are predictive of medullary thyroid cancer only 13% of the time (5, 6, 7, 8, 9, 14). Benign disease (e.g. follicular adenoma, thyroiditis, Graves disease, Hashimotos thyroiditis, and goiter), differentiated thyroid cancer, and other malignancies comprise the remaining disorders. This calculation presumes that the true incidence of medullary cancer in all these series is, in fact, known. However, it is likely that all major series to date, under-report the true incidence of medullary cancer. Niccoli et al. (14) reported two cases of medullary microcarcinoma in patients with normal basal calcitonin levels. It is likely that in other series, in which basal calcitonin levels are used as a screen for more definitive testing with pentagastrin stimulation, similarly undetected and unreported cases of medullary thyroid cancer must also exist.
We have analyzed data from published reports and have noted a statistically significant correlation between tumor size and basal calcitonin level (P < 0.01). The correlation coefficient between pentagastrin-stimulated calcitonin levels and tumor size, and also the correlation coefficient for assessing the relationship between basal and stimulated calcitonin levels, were not statistically significant.
The results from the study by Karanikas et al. (10) are significant and broaden our understanding of populations in whom calcitonin levels may be abnormal. These data do, however, require additional analysis. Karanikas et al. (10) do not have surgical data on the only two patients in their series who had elevated pentagastrin-stimulated calcitonin levels who lacked demonstrable nodular disease. One of these patients had resolving DeQuervains thyroiditis. This patient had elevated basal calcitonin levels with abnormal, but not pathologically elevated, stimulated calcitonin levels on repeat testing. The only patient in the series who was diagnosed with medullary thyroid cancerindeed the only patient in the study with elevated calcitonin levels who went to surgerywas a patient who had a thyroid nodule in addition to Hashimotos thyroiditis. This patient (patient 1 in the Karanikas series) would likely have gone to surgery for thyroidectomy solely because of the presence of a thyroid nodule and pathologically elevated calcitonin levels. The presence of thyroiditis would have been clinically immaterial.
Although it is clinically interesting to observe that patients with both nodular and non-nodular thyroid diseases may have similarly elevated basal and stimulated calcitonin levels, we do not believe this single series makes a convincing enough argument for routine screening in all patients with Hashimotos thyroiditis. Although we appreciate the results described in the study by Karanikas et al. (10), we suggest that additional studies be performed to validate these results before broadly recommending calcitonin screening in patients with Hashimotos thyroiditis.
Calcitonin testing has clear value, but is not a panacea. Numerous patients from the above cited studies with negative thyroid FNAs (and other preoperative evaluation) have had medullary thyroid cancers surgically removed, and probably cured, based only on the presence of elevated calcitonin levels. However, despite the tremendous momentum this mode of testing has accumulated in multiple studies done outside the United States, no consensus yet exists to recommend a unified approach to interpreting calcitonin levels. Both basal and pentagastrin-stimulated calcitonin values have been interpreted in these studies with different assays of varying sensitivity, making direct comparison difficult. No definitive data have been published that clearly identify a threshold that defines a pathological elevation of calcitonin after pentagastrin stimulation. Because pentagastrin is not available in the United States, this issue becomes even more problematic. The literature defines few values for other stimulation techniques such as calcium-stimulated calcitonin levels and the ability of those tests to discriminate benign thyroid disease, C-cell hyperplasia and medullary cancer.
Especially because pentagastrin is not available in the United States, the issue of calcitonin testing in patients with thyroid disease remains, in our minds, controversial. It does not seem to us that the use of basal calcitonin levels in the routine screening of patients with nodular or non-nodular thyroid disease is warranted at present without the ability to use pentagastrin stimulation as a confirmatory test. Even if pentagastrin stimulation is available, another serious question persists. To what extent has the incidence of C-cell hyperplasia, perhaps incorrectly identified as actual medullary thyroid cancer or a certain precursor of medullary thyroid cancer, clouded the diagnostic ability of the pentagastrin stimulation test? Medullary thyroid cancer is a serious disease and, as a general principle, it is requisite to try to diagnose and treat it as early as possible. However, the problems of false positive calcitonin elevations, the costs of the additional tests and procedures used to follow-up an elevated calcitonin level, and the frequent vagaries of deciphering a stimulated calcitonin value make it difficult to conclude that serum calcitonin should be measured routinely in patients with any type of thyroid disease.
If a consultant is presented with a patient with nodular thyroid disease who has already had an elevated serum calcitonin level, what additional evaluation should be performed? There are no evidence-based guidelines, but our opinion at present is that the test should be repeated with a specific double sandwich assay. If it is still elevated, then additional studies including a thyroid sonogram and FNA should be considered. A baseline calcitonin level that is greater than 100 pg/ml is very suspicious for the presence of medullary thyroid cancer, and usually these patients will require surgery to make a diagnosis. If the basal calcitonin level is between 10 and 100 pg/ml, the recommendations for subsequent evaluation are less clear. There are multiple data sets to suggest that markedly abnormal stimulation tests with pentagastrin greatly clarify the diagnosis of patients in this gray zone. For less convincingly abnormal tests or if pentagastrin is not available, clinicians also have the option of closely monitoring basal calcitonin levels with a reliable double sandwich assay. A clinical history of severe diarrhea, palpitations, diaphoresis, headache, hypertension, or hypercalcemia may increase suspicion for the presence of a medullary thyroid carcinoma. Clinical findings such as Marfinoid habitus, neural crest-derived mucosal skin or gastrointestinal lesions, and a family history of other neuroendocrine tumors or medullary thyroid cancer might also suggest the presence of a familial medullary carcinoma syndrome. In difficult cases, examination of thyroid glands of first-degree relatives that are available and measurement of their serum calcitonin levels might be useful. Very rarely, determination of ret oncogene and performance of additional radiological studies (e.g. somatostatin and positron emission tomography scanning) might be helpful. In a recent study, Bihan et al. (18) postulate that calcitonin precursor proteins may provide a more sensitive assay substrate than mature calcitonin. Newer methods such as proteomics may also provide more definitive strategies for diagnosing this sometimes very elusive disease.
In this extremely complex and controversial area of calcitonin screening for patients with thyroid disease, many questions remain unanswered. What are the causes of most false positive serum calcitonin levels? How well defined and reproducible are the histological criteria for C-cell hyperplasia? In the context of a negative ret oncogene, how often does C-cell hyperplasia evolve into medullary thyroid cancer? Is it cost effective to use basal calcitonin levels to screen patients with thyroid nodules? Now, Karanikas et al. (10) have raised a new question: What is the mechanism and significance of an elevated calcitonin level in patients with Hashimotos thyroiditis? Well-designed prospective controlled studies are required to address these issues. Until that time, all recommendations will continue to be based on incomplete evidence.
Acknowledgments
We thank Dr. Robert Gagel and Ms. Jacqueline Ennis for assistance and advice.
Footnotes
Abbreviation: FNA, Fine-needle aspiration.
Received November 14, 2003.
Accepted November 14, 2003.
References