| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Divisions of Endocrinology, Diabetes and Metabolism, and Nuclear Medicine and Thyroid Cancer Unit (R.T.K.) and Department of Medicine (E.L.M.), The Ohio State University, Columbus, Ohio 43210; and Division of Medicine (E.L.M.), University of Florida, Shands Hospital, Gainesville, Florida 32610
Address all correspondence and requests for reprints to: Dr. Mazzaferri, Professor of Medicine, Division of Endocrinology, Shands Hospital, 1600 SW Archer Road, P.O. Box 100226, Gainesville, Florida 32610-0226. E-mail: mazz01{at}bellsouth.net.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective: Our objective was to determine the accuracy of a single rhTSH-Tg measurement over time.
Design and Setting: We conducted a prospective follow-up study at the University referral center.
Patients: A total of 107 differentiated thyroid carcinoma patients were stratified according to their initial rhTSH-Tg as follows: group 1 with Tg less than 0.5 (n = 68), group 2 with Tg of 0.6–2.0 (n = 19), and group 3 with Tg greater than 2 ng/ml (n = 20).
Intervention: Clinical evaluations were conducted over 0.9–5.2 yr as follows: Tg during thyroid hormone suppression (n = 27), after rhTSH (n = 59), and/or after thyroid hormone withdrawal (n = 15).
Main Outcome: Tumor was identified in one patient in each of groups 1 (1.6%) and 2 (5.5%), and 16 in group 3 (80%), comprising 19 tumor locations: 11 locoregional, two mediastinal, five lung, and one brain. Tumor was found in 81% with an initial or follow-up rhTSH-Tg greater than 2 ng/ml. TSH-stimulated Tg fell spontaneously to less than 0.5 ng/ml in 50% of group 2 and 5% of group 3 over 1.7–5.0 yr. The positive predictive value of the initial rhTSH-Tg greater than 2 ng/ml was 80%, and the negative predictive value was 98%. After retreatment, 100% of group 1, 74% of group 2, and 55% of group 3 had no evidence of tumor (P = 0.0001).
Conclusions: 1) A single rhTSH-Tg greater than 2 ng/ml predicts persistent tumor, although no value entirely excludes future recurrence. 2) Repeated TSH-stimulated studies are appropriate for patients at risk of recurrence, especially those with an rhTSH-Tg greater than 1 ng/ml. 3) A single rhTSH-Tg less than 0.5 ng/ml without Tg antibody has an approximately 98% likelihood of identifying patients completely free of tumor, a large group in which TSH suppression to less than 0.1 mIU/liter and frequent imaging and TSH-stimulated Tg testing are unnecessary.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Surveillance to detect persistent tumor is thus an important facet of management. Much has been done in recent years, both in the United States and Europe, (4, 5) to improve follow-up paradigms, which now rely heavily upon TSH-stimulated thyroglobulin (Tg) measurements made with recombinant human (rh)TSH (rhTSH-Tg) or after thyroid hormone withdrawal (THW) (Tg-off), and neck ultrasonography, posttreatment whole-body scans (RxWBS), and in selected cases, 18fluorodeoxyglucose (18FDG) positron emission tomography (PET)/computed tomography (CT) scans.
It is important that tests for thyroid cancer have a high negative predictive value (NPV) because this reassures both the patient and physician that the tumor has been eradicated, which reduces the cost and complexity of follow-up and permits the levothyroxine dosage to be lowered, avoiding the complications of subclinical thyrotoxicosis (4). Schlumberger and colleagues (6) found a high NPV (99%) of serum Tg-off and also showed that Tg was detectable in 15% of DTC patients who had no other evidence of disease, including negative diagnostic whole-body scans (DxWBS). The importance of TSH-stimulated serum Tg, whether by Tg-off or rhTSH-Tg, was quickly confirmed (4, 7, 8). Subsequent studies have consistently shown that ultrasonography and TSH-stimulated serum Tg levels together have the highest sensitivity for detecting residual tumor in low-risk patients (9, 10, 11, 12, 13), affirming that the DxWBS has little place in their follow-up management, and that an undetectable Tg measured during TSH suppression (Tg-on) is unreliable (7). Recent consensus papers from the United States (4) and Europe (5) strongly endorse these concepts.
Still, the positive predictive value (PPV) of the initial Tg-off is low, often ranging from 42–53% for Tg cutoffs, respectively, of 5 and 10 ng/ml (9). Likewise, the PPV of an rhTSH-Tg level above 2 ng/ml is around 50% in many studies (7, 14, 15), suggesting that it is sometimes as accurate as a coin toss, potentially resulting in unnecessary additional testing and treatment of many patients. In fact, about one fifth of patients thought to be free of disease have rhTSH-Tg levels above this cutoff (4). Serial measurements of TSH-stimulated Tg, however, increase the sensitivity of the test, which in one study increased the PPV of Tg-off to 83% when the stimulated Tg increased over time (9). It would be an improvement if one rhTSH-Tg test would accurately establish the ultimate outcome in the majority of patients.
Any improvement in follow-up paradigms for patients with DTC will have a major economic and human impact because the prevalence of thyroid carcinoma is so high, estimated at about 330,000 in the United States and 200,000 in Europe. Most of these individuals have DTC, and perhaps as many as 20% have residual disease after treatment of their primary tumor. Improvements on test accuracy and surveillance paradigms are expected to improve prognosis by detecting tumor early in its course at a low stage.
The aim of this study was to assess the accuracy of an initial serum rhTSH-Tg measurement in a cohort of 107 patients with DTC first reported in 2002 (7). We report the ongoing development and treatment of tumor in some and the spontaneous fall of stimulated serum Tg in others in this cohort with follow-up ranging from 0.9–5.2 yr after the initial study.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The Ohio State University Biomedical Sciences and The Arthur G. James Cancer Hospital Institutional Review Boards approved our ongoing follow-up study of patients with thyroid cancer, including this follow-up study of 107 consecutively evaluated patients with papillary, follicular, or Hürthle cell carcinoma who were initially reported in 2002 (7) after they underwent evaluation for persistent tumor using rhTSH, shortly after it had been approved for clinical use. Patients were tested between January 1999 and March 2001 if they had no anti-Tg antibodies (TgAb) and had been free of disease on the basis of clinical examination and had one or more undetectable (<0.5 ng/ml) or low (0.6–1 ng/ml) serum Tg-on measurements. All had undergone a 4-mCi 131I diagnostic DxWBS and an RxWBS after initial 131I ablation of thyroid remnants or after treatment of recurrences, and had chest x-rays, and in some cases, ultrasonography, CT, MRI, or 18FDG-PET imaging during the initial study.
Eighty-eight (82%) were females and 19 (18%) were males, who at the time of initial surgery ranged in age from 10.9–85.3 yr (median, 36.3 yr). They had undergone ostensibly successful initial therapy, although three patients were found to have residual tumor that was thought to have been successfully treated before the first rhTSH evaluation. Patients were given rhTSH, 0.9 mg im on 2 consecutive days followed in the first study by 4 mCi 131I 24 h later and serum Tg measurement and DxWBS on the fifth day (72 h after the last rhTSH injection), as previously described (16). Time from initial surgery to the first round of rhTSH tests was 10 months to 35 yr [median, 3.3 yr; mean, 6.9 ± 0.4 (SEM) yr]. Average patient age at the time of the initial evaluation was 44.0 ± 1.3 yr and ranged from 14.4–88.2 yr (median, 43.2 yr).
Initial therapy
Total or near-total thyroidectomy was performed on all 107 patients, sometimes (36%) after completion thyroidectomy. Lymph node surgery was simple excision in 19 (18%), ipsilateral modified neck dissection in seven (7%), and bilateral modified neck dissection in three (3%) patients and no lymph node surgery in 78 (72%) patients. After surgery, all were treated with 131I ranging in amounts from 26–228 mCi (median, 96 mCi). Cumulative 131I administered by the time of the initial report in 2002 ranged from 26–800 mCi (median, 105 mCi).
Tumor characteristics, stage, and tumor discovery after rhTSH testing
The tumors were mainly classic papillary thyroid carcinomas (PTC) (71%), but some were follicular variant (9%), tall cell variant (2%), and Hürthle cell variant (1%) PTC and follicular (11%) or Hürthle cell (6%) carcinomas (Table 1
). More than half of the patients (66) had tumors that were at high risk of recurrence (T3, T4, or N1 or M1).
1
|
). Tumor was identified in three others after initial staging and 131I thyroid remnant ablation (patients 4, 39, and 90). After additional therapy, they also were thought to be free of disease with undetectable serum THST-Tg levels when the first rhTSH study was performed (Table 2).
|
), whereas two patients (36 and 51) with high Tg-off serum Tg levels (greater than 10 ng/ml) were considered to have tumor despite imaging studies that failed to locate its site. Thus, metastatic tumor was disclosed by rhTSH testing in 11 of these 107 patients (10%) who were otherwise clinically free of disease.
|
Follow-up studies began in June 2001 and continued until September 2004. Patients were categorized as having no evidence of disease (NED) if neck ultrasonography was negative and they had either 1) one or more TSH-stimulated (rhTSH-Tg or Tg-off) serum Tg levels less than 0.5 ng/ml after the first rhTSH study or 2) at least two subsequent Tg-on levels less than 0.5 ng/ml 6 or more months apart after the first rhTSH study. Persistent tumor was identified by fine-needle aspiration cytology or in surgical specimens or by 131I WBS showing 131I uptake outside the thyroid bed. Chest x-ray, CT, MRI, 131I DxWBS and/or RxWBS, or 18FDG-PET scans were performed as indicated, usually when the Tg was elevated and neck ultrasonography was negative. Tg levels reported in this follow-up study are those obtained at the time the patient’s tumor status changed or at the end of the study when it remained undetectable.
Tg measurement
The average serum Tg value is reported from duplicate measurements made in our laboratories using the Nichols chemiluminescence immunometric assay (Nichols Institute Diagnostics, San Diego, CA; catalog no. 60-4240; analytical sensitivity, 0.07 ng/ml; functional sensitivity, 0.5 ng/ml). The coefficients of variation for three controls (1.49–2.29 and 18.10–21.90 ng/ml for the kit controls and 3.93–6.01 ng/ml for the serum pool control) were 10.6, 4.8, and 10.5%, respectively, and 24.8% for a Tg level of 0.44 ng/ml in a pooled serum control. The normal range is 3.2–57.7 ng/ml, with a mean of 13.5 ng/ml. To exclude a hook effect, all results were confirmed by measurement of 1:10 dilution aliquots. Serum samples were frozen and reassayed with the latest Tg sample if deemed necessary (7).
TgAb measurement
TgAb was quantitated with each Tg measurement using the Nichols chemiluminescence Tg antibody kit (catalog no. 60-4185; sensitivity, 2.0 U/ml). Using local pooled control standards, the coefficient of variation for a TgAb level of 1.23 U/ml (below the assay sensitivity) is 16.7%; for antibody levels ranging from 3.5–34.5 U/ml, the coefficient of variation ranges from 10.5–13.4%. A negative test is an antibody level less than 2 U/ml.
Neck ultrasonography
Neck ultrasonography was performed by R.T.K. using a GE Logiq 200 
with 7.5-MHz linear probe inspecting the superior mediastinum and the bilateral central and the bilateral lateral neck compartments.
Statistics
All results are reported as mean ± SE. Percentages in the text are rounded to the nearest integer. Tg levels clinically reported as less than 0.5 ng/ml are considered as 0.5 ng/ml for statistical purposes. Data analysis was done with Stata Statistical Software release 4.0 (Stata Corp., College Station, TX). Fisher’s exact test was applied to categorical data. ANOVA was used to detect differences among group means. Stepwise logistic regression analysis was used to compare variables predicting recurrence. A two-tailed P value less than 5% was considered statistically significant. The rhTSH-Tg data from the original study were used to stratify patients into three groups as described below. During follow-up, the Tg values that identified the patient’s current tumor status are reported. The accuracy of serum Tg testing was calculated from the rhTSH-Tg data from the first study.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
). Six patients were lost to follow-up shortly after the first rhTSH study. Patients were evaluated an average of 3.3 ± 0.9 yr (median, 3.2; range, 0.9–5.2 yr) after their first rhTSH study. Time from initial surgery to the current evaluation ranged from 1.8–39 yr (median, 6.7; mean, 8.7 ± 0.6 yr). Individual relapse time is shown in Tables 3
and 4. The average age of patients when their current tumor status was determined was 47.2 ± 0.1 yr and ranged from 23.7–92.6 yr (median, 46.6 yr).
|
|
In the first rhTSH study, the baseline serum Tg-on was less than 0.5 ng/ml in 102 (95%) of 107 patients and was 0.6 ng/ml in four patients and 1.0 ng/ml in one patient. In the present study, Tg-on was less than 0.5 ng/ml in 95 (94%) of the patients, and ranged from 0.9–6.2 ng/ml in six others, all in group 3.
Tg responses to rhTSH during the first rhTSH study
To compare the results of the current follow-up study with the original one, patients were stratified into three groups according to their serum rhTSH-Tg responses in the original study: group 1 with Tg less than 0.5 (68 patients), group 2 with Tg from 0.6–2.0 ng/ml (19 patients), and group 3 with Tg greater than 2 ng/ml (20 patients).
Current Tg levels in each study group
Tg less than 0.5 ng/ml (group 1). In the initial study, none of the 68 patients in this group had evidence of tumor on physical examination, chest x-ray, DxWBS, and neck ultrasonography. Five were lost to follow-up without evidence of disease, leaving 63 (93%) of group 1 patients for additional follow-up studies.
During the second follow-up period, 32 patients (47%) in group 1 underwent one or more rhTSH-Tg studies, and 10 (15%) had one or more Tg-off measurements resulting in a mean TSH of 38.19 ± 1.2 mIU/liter. Twenty-one patients (31%) had Tg-on measurements over an average of 2.9 ± 0.2 yr after the first negative rhTSH test had been performed. TSH-stimulated and Tg-on levels in this group remained undetectable (<0.5 ng/ml) during follow-up. Still, one patient (no. 64) developed anti-Tg antibodies at the time of follow-up testing when her rhTSH-Tg was less than 0.5 ng/ml. This prompted further neck ultrasonography that detected tumor, and she underwent modified neck dissection that removed six malignant lymph nodes from the lateral neck (Table 4). Although she currently has no detectable tumor, Tg antibodies remain detectable postoperatively. No other patients in group 1 were found to have disease during follow-up. Thus, 1.5% of the original cohort of 68 group 1 patients had persistent tumor despite their original rhTSH-Tg of no more than 0.5 ng/ml without interfering antibodies about 3 yr earlier.
Tg 0.6–2.0 ng/ml (group 2). None of the 19 patients in this group had evidence of tumor at the conclusion of the first study, when all had a baseline serum Tg of less than 0.5 ng/ml that increased to levels ranging from 0.6–2.0 ng/ml in response to rhTSH. One patient in this group was lost to follow-up with NED, leaving 18 patients for additional evaluation.
During the follow-up evaluation, 13 patients (68%) had one or more rhTSH-Tg tests, and one (5%) underwent Tg-off testing. Four (21%) had only Tg-on measurements performed, all of which remained undetectable (<0.5 ng/ml) throughout the follow-up period averaging 2.9 ± 0.05 yr (range, 2.8–3.0 yr) after the first rhTSH test.
The rhTSH-Tg or Tg-off levels became undetectable (<0.5 ng/ml) without further therapy in nine (50%) of the group 2 patients (Fig. 2). In four others, TSH-stimulated Tg levels increased to 0.7, 0.8, 1.1, and 1.5 ng/ml, but no tumor was found. In another patient (no. 52) with Tg-on of 0.5 ng/ml, the rhTSH-Tg increased to 3.4 ng/ml (Fig. 2 and Table 4). This 28.6-yr-old woman had a Tg-on of less than 0.5 ng/ml and an rhTSH-Tg of 0.6 ng/ml 3.5 yr before the present test. After a positive neck ultrasonography, malignant lymph nodes were excised from her left central neck compartment, but her early follow-up rhTSH-Tg was 1.5 ng/ml, and 10 months later more malignant lymph nodes were resected from the left central neck compartment after repeat positive neck ultrasonography. Five and one half months after the last operation her Tg-on was less than 0.5 ng/ml and rhTSH-Tg was 3.4 ng/ml. Her last neck ultrasound and chest CT examinations were negative, and her status is classified as NED with a persistently elevated serum Tg level (Table 5). None of the other patients in this group have had evidence of tumor. Thus, 5.3% of the original 19 group 2 patients had persistent tumor detected within approximately 3 yr of follow-up after initial rhTSH-Tg values of 0.6–2.0 ng/ml.
|
|
). Nine (45%) were found to have persistent tumor identified by fine-needle cytology, surgical pathology, anatomic imaging, or RxWBS (Table 3). Tumor was identified in the lung in four patients and in lymph nodes in five (three in central neck compartments, one in both lateral cervical compartments, and one in the mediastinum), but the tumor site could not be identified in two others (nos. 36 and 51) with high serum rhTSH-Tg levels (respectively, 4.6 and 7.0 ng/ml) and high Tg-off levels (respectively, 28.5 and 70.6 ng/ml). Nine (8%) others had rhTSH-Tg levels above 2 ng/ml, but tumor could not be found on physical examination, chest x-ray, DxWBS, and ultrasonography in all the patients and chest CT (n = 5), neck CT (n = 4), neck MRI (n = 1), brain MRI (n = 1), PET (n = 2), and RxWBS (n = 2) in some. One patient had significant thyroid bed uptake on the RxWBS. Tumor thus was identified in only nine patients in group 3 during the first study, whereas it was not found in 11 others (55%).
|
and 5). The other patient (no. 80) elected to be followed without further investigation despite strong suspicion of persistent disease (Fig. 1). Tumor was found in seven more patients during follow-up (Table 4), five with cervical lymph node metastases, one with mediastinal metastases, and one with lung metastases. Two of the patients (nos. 36 and 51) with lymph node metastases are the ones mentioned above with very high Tg levels in whom the tumor site could not be identified in the first study. Thus, tumor was identified in a total of 16 (80%) of 20 group 3 patients during the first and second studies (Tables 3 and 4). One other patient (no. 90) who had tumor identified in the first rhTSH study (Table 3) had a new tumor location identified in the brain during the follow-up study, resulting in 17 tumor locations among the 16 group 3 patients with tumor. Thus, 80% of the 20 group 3 patients had persistent tumor identified within approximately 5 yr from their first rhTSH-Tg value of greater than 2.0 ng/ml. Tumor has not been identified in four group 3 patients. Tumor locations detected during follow-up
A total of 18 (17%) of 107 patients had persistent tumor, which was identified in all but one case by rhTSH-Tg greater than 2 ng/ml (Tables 3 and 4). Tumor was found in 81% (17 of 21 patients) with a rhTSH-Tg level greater than 2 ng/ml, including one group 2 patient (no. 52) with a rhTSH-Tg of 3.4 ng/ml in the follow-up study and 16 group 3 patients. Tumor was found by ultrasonography alone in one group 1 patient (no. 64) with an undetectable Tg (<0.5 ng/ml), both during Tg-on and rhTSH-Tg, who developed recurrent TgAb at the time of follow-up testing.
Nineteen tumor locations were identified among 18 patients, 11 (58%) in the neck (nine in the central and two in the lateral compartments), five (26%) in the lung, two (11%) in the mediastinum, and one in the brain (Tables 3 and 4).
Spontaneous decline in serum TSH-stimulated Tg to undetectable levels
After initial therapy, 10 patients had a spontaneous decline in TSH-stimulated serum Tg levels (eight rhTSH-Tg and two Tg-off) representing 9% of the 107 patients in the initial study group (nine in group 2 and one in group 3). This occurred in 50% of group 2 and only 5% of group 3 patients (Fisher’s exact test, P < 0.0001).
The median time for the Tg to decline to less than 0.5 ng/ml was 4.2 yr from the first rhTSH test (mean, 3.7 ± 0.36; range, 1.7–5.0 yr). The tumor classification was 20% T1, 50% T2, 10% T3, and 20% T4 in those in whom TSH-Tg spontaneously fell after initial therapy, which was not statistically different from the tumor classification in the rest of the study group (P = 0.076, Fisher’s exact test; Table 1). Similarly, the nodal classification was 80% N0 and 20% N1 in those in whom the TSH-Tg spontaneously declined, which was not statistically different from the nodal classification in the rest of the study group (P = 0.79, Fisher’s exact test; Table 1).
Tumor stage and outcome
There were no differences in the tumor node metastasis (TNM) classification in the three groups (ANOVA, P = 0.61; Table 1) or in patient age, gender, and TNM status by stepwise regression analysis (P = 0.4). Thus relapse could not be anticipated by the presenting manifestations of the tumor, even in those with aggressive disease (Table 2).
Therapeutic interventions and outcome
Thirty-three therapeutic interventions were performed after initial treatment. Therapy for the relapses included 14 neck operations, 17 administrations of 131I, one external beam radiotherapy, and one experimental therapy in a clinical trial. One patient with progressive disease declined experimental treatment. At last evaluation, 93 patients (87%) had no evidence of tumor, 100% of group 1, 74% of group 2, and 55% of group 3, differences that are highly significant (ANOVA, P = 0.0001). All but one of the group 2 patients had NED, although a few had slight elevations of TSH-stimulated serum Tg levels ranging from 0.7–1.5 ng/ml (Fig. 2 and Table 5) without ultrasonographic or other evidence of tumor. Thus, about 25% of group 2 and group 3 had detectable serum rhTSH-Tg levels with negative imaging studies (Tables 4 and 5). Three group 3 patients (15%) with lung or mediastinal metastases have persistent but stable tumor, whereas another two group 3 patients have progressive disease that may be fatal (Tables 3 and 5). Patients who underwent treatment and became free of disease were 3-fold as likely to have become so after additional surgery compared with additional 131I treatment (Tables 3 and 4).
Accuracy of rhTSH-stimulated serum Tg measurements
A positive test is a serum rhTSH-Tg greater than 2 ng/ml in the first rhTSH study in a patient subsequently found to have one or more tumor foci. A negative test is described in the Patients and Methods section. According to these criteria, the PPV of an rhTSH-Tg greater than 2 ng/ml was 80%, NPV 98%, sensitivity 89%, and specificity 96% (Fig. 4). Without the patient with TgAb who developed TgAb shortly after the first rhTSH testing, the NPV was 99%.
|
In this series, 32 patients (30%) underwent their first rhTSH-Tg shortly after initial surgery and 75 others (70%) had their first rhTSH testing up to several decades later. The mean time for the first rhTSH-Tg test after initial therapy in groups 1, 2, and 3 was, respectively, 5.5 ± 0.7 yr with a 95% confidence interval (CI) of 4.0–6.9, 4.9 ± 1.3 yr (95% CI 2.3–7.6), and 5.5 ± 1.3 yr (95% CI 2.7–8.3) (P = 0.9 by ANOVA). Stepwise regression of recurrence/persistent tumor (dependent variable) on age, first Tg-rhTSH, first Tg-on, TNM status, and time lapse from first surgery to initial rhTSH testing showed that only the first serum rhTSH-Tg measurement predicted recurrence (P < 0.0001). When the time interval between rhTSH-Tg testing and initial therapy was dichotomously stratified into two groups, 2–24 months and 25 months to 37 yr after initial therapy, the recurrence rates in the two groups, respectively, were 19 and 15% (P = 0.6 by Fisher’s exact test). The accuracy of rhTSH-Tg testing was thus similar throughout the extended follow-up period during which patients in this study were first tested with rhTSH.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
There are data showing that a delay in primary treatment has a large impact on outcome, and it thus seems likely that an adverse outcome may result from a delay in the treatment of unrecognized disease that persists after initial management. We previously found that a delay in therapy of more than 12 months doubled the thyroid cancer mortality rate, which at 30 yr was 13% after a delay compared with 6% for all others (P < 0.001) (18). Multivariate analysis showed that a delay in primary treatment increased the probability of cancer death 2.4-fold (1). In another study, multivariate analysis found that patients who underwent completion thyroidectomy within 6 months of their primary operation developed significantly fewer lymph node and hematogenous recurrences and survived significantly longer than those in whom the second operation was delayed longer than 6 months (19). Beierwaltes et al. (20) found that patients with PTC who had NED were more likely to have been treated with 131I within 3 months of their thyroidectomy.
Late recognition of tumor during follow-up has historically been a common problem, with as many as 20% of patients experiencing tumor relapse, often decades after total thyroidectomy and 131I remnant ablation despite careful surveillance with older testing tools (1). In the 1980s and 1990s, the majority of tumors found after initial therapy fell below the detection limits of relatively insensitive tests, including physical examination, chest x-ray, inaccurate Tg assays, and 131I DxWBS. As a result, serious metastatic tumor was often first detected late in the course of the disease. In our cohort, 15% of the locoregional recurrences and 25% of distant metastases were identified more than two decades, and in some cases as late as 40 yr, after initial therapy (1). Distant metastases, mostly to lung, comprised 32% of the recurrences and accounted for almost half the cancer deaths, whereas 68% of the relapses were locoregional tumors, causing 25% of all the cancer deaths (1). Thirty-year cancer mortality rates were twice as high with neck soft tissue tumor recurrence (30%) compared with cervical lymph nodes or tumor in the contralateral thyroid (16%) (1). A recent study from France (21) underscores the seriousness of advanced locoregional recurrence, more than half (60%) of which were discovered by physical examination. Ten-year survival rates in this group of 172 patients with locoregional recurrence were 49.1% for all patients, 89.3% for those less than age 45 yr at the time of initial therapy, and only 32.1% for patients older than 45 yr (21).
Similar observations of adverse outcomes associated with more advanced disease, presumably because of delayed diagnosis, have been made in patients with lung metastases. In one study (22), complete remission and 10-yr survival rates, respectively, were 96 and 100% when the tumor was found only by a high Tg and RxWBS but fell to 83 and 91% when the tumor was found with normal chest x-ray and a positive DxWBS, 53 and 63% when it was manifest as micronodules on chest x-ray, and 14 and 11% when x-ray showed macronodules.
With newer surveillance paradigms and more sensitive serum Tg assays, it is often difficult to identify tumor promptly in patients with high serum Tg levels and to differentiate those in whom tumor will be identified from others in whom the Tg will spontaneously decline over time. Schlumberger’s group (6) studied 256 patients who did not have uptake outside the thyroid bed on the initial RxWBS after thyroid remnant ablation and found that only 33% of 15 patients with a Tg-off above 10 ng/ml had tumor. The authors attributed this low number of tumors to the relatively short follow-up of about 1 yr and the slow growth rate of most DTCs. Another later study (9) of the same patient cohort, in which nearly 15% were found to have an initial Tg-off above 1 ng/ml, found that only 24% had tumor 3–117 months later. After a median of 52 months, 31% had an undetectable (<1 ng/ml) serum Tg-off level. The PPV of the initial Tg-off above 5 ng/ml and 10 ng/ml, respectively, was only 42 and 53% and was only 50% at the time of recurrence or subsequent follow-up. This low PPV was thought to be a result of the relatively low recurrence rate in this group of patients despite fairly long follow-up and of the subsequent spontaneous decline of serum Tg levels without additional therapy in 38% of patients, which has been observed in other studies (23) and sometimes occurs 2 or more years after therapy (12).
In our study, patients were first tested with rhTSH-Tg within a few months to many years after undergoing initial surgery. Persistent tumor was identified in 17% of our patients by a single rhTSH-Tg above 2 ng/ml, regardless of where they were in the follow-up cycle. Residual tumor was found by ultrasonography alone in one patient with an undetectable rhTSH-Tg when she developed TgAb that were not present during the initial testing. A spontaneous decline of TSH-stimulated serum Tg to undetectable levels occurred in only 10 of 39 (26%) of our group 2 or group 3 patients, which was significantly more likely to occur (50%) when the initial rhTSH-Tg was not higher than 2 ng/ml compared with only 5% when it was over 2 ng/ml (P < 0.0001). The median time for the Tg to decline to less than 0.5 ng/ml was 4.2 yr from the initial rhTSH test, with a range of about 6 months to 5.0 yr. Neither tumor recurrence nor a spontaneous decline in serum Tg levels could be predicted by the TNM status of our patients.
After 31 secondary therapeutic interventions, 87% of our patients had no evidence of tumor, an outcome that was closely related to the initial rhTSH-Tg level. After therapy for persistent tumor, there was no evidence of disease in 100% of group 1, 74% of group 2, and 55% of group 3 (P = 0.0001). Stimulated serum Tg levels remained measurable, but imaging studies remained negative in 26% of group 2 and 20% of group 3 patients after treatment of persistent disease. Among the group 3 patients with tumor, two have lung and one has mediastinal metastases that are stable or improved, whereas two others have progressive disease, including the development of brain metastases in one person.
A limitation of the present study is that not all the patients underwent TSH-stimulated serum Tg measurements during follow-up and there might be patients with early tumor that have not yet been identified despite ultrasound and repeated negative serum Tg-on testing. If we limit our analysis to patients who underwent rhTSH-Tg or Tg-off testing in follow-up, then tumor was found in 2% of group 1 (one of 42), and 7% of group 2 (one of 14) patients, values similar to those reported when all patients are included. Another limitation of our study is the median follow-up of only 3.2 yr so that all patients destined to manifest metastases still may not have been identified. Indeed, one group 1 patient (no. 67) not stimulated with rhTSH in the second study has since the close of this study demonstrated an rhTSH-Tg of 2.7 ng/ml and is currently undergoing further evaluation.
Metastases were found in 17% of our patients, all but two of whom had an rhTSH-Tg level greater than 2 ng/ml during the first study. It is probable that the one group 1 patient (no. 64) with lymph node metastases and an undetectable rhTSH-Tg had tumor at the time of the first rhTSH evaluation and developed TgAb 34 months later that prompted repeat neck ultrasonography examinations. This case emphasizes several points. First, small lymph node metastases detected by ultrasonography are often associated with undetectable Tg-on levels that occasionally fail to rise with rhTSH or THW (10, 11, 12, 13, 24). Second, TgAb seriously interfere with Tg immunometric assays, even at very low levels (25). Third, recurrent TgAb serve as a warning of recurrent disease. Moreover, a one-time undetectable TSH-stimulated serum Tg does not always assure a permanent long-term remission.
The follow-up paradigms that have evolved using neck ultrasonography and rhTSH-Tg or Tg-off measurements have mainly targeted patients at low risk for recurrence. Still, this excludes the 15% or more high-risk patients in most studies (4, 5). More than half of our patients were at high risk of recurrence according to the TNM status or tumor histology, although neither accurately predicted persistent tumor. This emphasizes the value of an RxWBS after remnant ablation and early restaging of tumor with rhTSH-stimulated Tg levels and ultrasonography, an approach advocated by Pacini, Schlumberger, and ourselves (4, 5, 6, 7, 8, 9).
The PPV of an initial rhTSH-Tg greater than 2 ng/ml in this study was 80%, and the NPV was 98%. The sensitivity and specificity, respectively, were 89 and 96%. If the one patient with TgAb at follow-up is omitted from the analysis, the NPV of rhTSH-Tg becomes 99%. Measured during thyroid hormone suppression of TSH, the Tg level had a sensitivity of only 22% and a NPV of 86%, which increases only to 87% if the patient with TgAb is omitted from the analysis (Fig. 4). This suggests that a single rhTSH-Tg test, even in high-risk patients, is likely to identify those with persistent tumor, even years after initial therapy.
Follow-up of patients with a relatively small TSH-stimulated rise in Tg to no higher than 2 ng/ml is not usually a serious matter but requires follow-up with neck ultrasonography and rhTSH-Tg or Tg-off testing, mainly to determine whether the serum Tg levels are rising, because such a patient may have persistent tumor. The frequency of such testing remains uncertain but probably should be done annually until the direction and magnitude of the rhTSH-Tg levels are determined.
On the other hand, a single serum rhTSH-Tg that rises above 2 ng/ml is a more serious matter that requires more evaluation. An aggressive approach to these patients is supported by the fact than more than half became free of disease after treatment and at least short-term follow-up.
An undetectable (<0.5 ng/ml) rhTSH-Tg in a patients without serum TgAb has a 98% likelihood of identifying patients who are free of tumor, a large patient group in which TSH suppression to low levels appears unnecessary. Annual Tg-on measurement is recommended. A few of these patients will eventually demonstrate metastases, typically in small cervical lymph nodes. The optimal role and frequency of neck ultrasonography and/or rhTSH-Tg testing in these patients is not known, but annual testing does not seem warranted.
|
Footnotes |
|---|
Abbreviations: CI, Confidence interval; CT, computed tomography; DTC, differentiated thyroid carcinoma; DxWBS, diagnostic whole-body scan; FDG, fluorodeoxyglucose; MRI, magnetic resonance imaging; NED, no evidence of disease; NPV, negative predictive value; PET, positron emission tomography; PPV, positive predictive value; PTC, papillary thyroid carcinoma; rh, recombinant human; RxWBS, posttreatment whole-body scan; Tg, thyroglobulin; TgAb, anti-Tg antibodies; Tg-off, after thyroid hormone withdrawal; Tg-on, during TSH suppression; THW, thyroid hormone withdrawal; TNM, tumor node metastasis.
1 The tumors were staged according to the fifth edition and not the sixth edition of the American Joint Committee on Cancer Manual for Staging of Carcinoma, Springer Verlag, 2003, to maintain consistency with the first publication of this cohort. 
Received March 8, 2005.
Accepted June 13, 2005.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. Elisei, M. Schlumberger, A. Driedger, C. Reiners, R. T. Kloos, S. I. Sherman, B. Haugen, C. Corone, E. Molinaro, L. Grasso, et al. Follow-Up of Low-Risk Differentiated Thyroid Cancer Patients Who Underwent Radioiodine Ablation of Postsurgical Thyroid Remnants after Either Recombinant Human Thyrotropin or Thyroid Hormone Withdrawal J. Clin. Endocrinol. Metab., November 1, 2009; 94(11): 4171 - 4179. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Pacini, M. G. Castagna, L. Brilli, G. Pentheroudakis, and On behalf of the ESMO Guidelines Working Group Differentiated thyroid cancer: ESMO Clinical Recommendations for diagnosis, treatment and follow-up Ann. Onc., May 1, 2009; 20(suppl_4): iv143 - iv146. [Full Text] [PDF]
|
|
|
|
 |
|
 A. N. Hoofnagle, J. O. Becker, M. H. Wener, and J. W. Heinecke Quantification of Thyroglobulin, a Low-Abundance Serum Protein, by Immunoaffinity Peptide Enrichment and Tandem Mass Spectrometry Clin. Chem., November 1, 2008; 54(11): 1796 - 1804. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. T. Kloos Approach to the Patient with a Positive Serum Thyroglobulin and a Negative Radioiodine Scan after Initial Therapy for Differentiated Thyroid Cancer J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1519 - 1525. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. Castagna, L. Brilli, T. Pilli, A. Montanaro, C. Cipri, C. Fioravanti, F. Sestini, M. Capezzone, and F. Pacini Limited Value of Repeat Recombinant Human Thyrotropin (rhTSH)-Stimulated Thyroglobulin Testing in Differentiated Thyroid Carcinoma Patients with Previous Negative rhTSH-Stimulated Thyroglobulin and Undetectable Basal Serum Thyroglobulin Levels J. Clin. Endocrinol. Metab., January 1, 2008; 93(1): 76 - 81. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. C. Smallridge, S. E. Meek, M. A. Morgan, G. S. Gates, T. P. Fox, S. Grebe, and V. Fatourechi Monitoring Thyroglobulin in a Sensitive Immunoassay Has Comparable Sensitivity to Recombinant Human TSH-Stimulated Thyroglobulin in Follow-Up of Thyroid Cancer Patients J. Clin. Endocrinol. Metab., January 1, 2007; 92(1): 82 - 87. [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 |
, Patient with TgAb that became detectable at the time of follow-up testing.