| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Institutes of Pathology (C.A., P.G., J.A.L.), Clinical Chemistry (C.A.), Internal Medicine (P.N.), and Nuclear Medicine (H.R.), Inselspital, Institute of Social and Preventive Medicine (C.M.), University of Bern, CH-3010 Bern, Switzerland; and Institute of Nuclear Medicine (C.A.), Clinique Ste. Therese, L-2763 Luxembourg, Luxembourg
Address all correspondence and requests for reprints to: Claudine Als, M.D., Division of Clinical Chemistry, Inselspital, University of Bern, CH-3010 Bern, Switzerland. E-mail: . claudine.als{at}insel.ch
Abstract
Among patients with differentiated thyroid carcinoma (diffTCa), the rare hyperfunctioning or toxic TCa (ToxTCa) was diagnosed when scintigraphic hot thyroid areas were attributable only to diffTCa (diameter >1 cm by pathological examination) and/or total thyroidectomy failed to induce hypothyroidism. Of 924 cases of all TCa (papillary diffTCa 47.3%, follicular diffTCa 44.2%, others 8.5%), 19 had ToxTCa (2.1%, 15 of 19 follicular, 4 of 19 papillary, P = 0.008). These received a more intensive radioiodine therapy (average cumulated 131I activities 21.8 vs. 15.2 GBq, P < 0.01). Five-year survival rates for ToxTCa (n = 19, 56%) and diffTCa (n = 545, 94.5%) differed [hazard ratio 4.8, 95% confidence interval (CI) 2.8–8.1, P = 0.001]. However, the differences were attenuated by matching ToxTCa and diffTCa (n = 57, 5-yr survival rate 74%) for age, sex, and histopathologic type (hazard ratio 2.1, 95% CI 1.13–3.9, P = 0.02). Correcting statistically for M1 against M0 stage distribution resulted in a further reduction of the hazard ratio (hazard ratio 1.8, 95% CI 0.93–3.48, P = 0.08). An M1 stage is an important prognostic factor in ToxTCa patients. Thus, ToxTCa, treated with higher activities of 131I, has a survival prognosis close to that of matched diffTCa cases, both groups consisting mainly (79%) of follicular subtypes.
DIFFERENTIATED THYROID CARCINOMA (diffTCa) is generally associated with euthyroidism. DiffTCa and thyrotoxicosis may coexist, originating from focally separate or identical parenchymal thyroidal volumes. If a diffTCa cell mass occupies a scintigraphically hot, functionally autonomous thyroid nodule (AFTN) and if thyrotoxicosis is attributable exclusively to the diffTCa cell mass, as shown by clinical-histopathologic correlation, a toxic thyroid carcinoma (ToxTCa) is diagnosed.
Because of mutations, deletions, gene duplications, RNA splicing errors, or aberrant expressions, the molecular machinery for iodide organification and processing of thyroid hormones may become less efficient in diffTCa, compared with nonmalignant thyroid cells. As shown in a case report of a diffTCa with lung metastases, this might be due to activation of oncogenes (i.e. ras) and the subsequent suppression of the differentiated functions (1 1A, 2). Reduced uptake of radioiodine because of low expression of the sodium/iodide symporter may be the consequence of an early step toward cancer transformation (3), although the expression pattern of the symporter is variable in diffTCa: It may be increased, decreased, or absent (4, 5). Functionality of the symporter with increased iodine uptake has recently been shown to be chemotherapeutically stimulable (6).
Conversely, functionally autonomous ToxTCa causing thyrotoxicosis is expected to represent a highly differentiated subgroup of diffTCa (7), with probably preserved molecular functions for processing of thyroid hormones. Somatic mutations of the TSH receptor genes play an important pathogenetic role for solitary AFTN (8, 9, 10) and also for at least a subset of ToxTCa (1 1A, 11, 12, 13). Whether such mutations in the Gs
protein coupled receptor are only of functional consequence or whether they are responsible for malignant transformation remains currently unclear (10, 14, 15). These mutations lead to constitutive activation of the cAMP cascade, which stimulates both growth and function. The resulting AFTN, peripherally euthyroid, clinically irrelevant at beginning, possibly have similar prevalences anywhere in the world. In an iodine-deficient environment, however, mutated follicular cells present a differential advantage of growth and function over unaffected cells (16). If iodine supply is thereafter increased after a critical AFTN volume has been reached, thyrotoxicosis may result. This is possibly the reason why goiter and thyrotoxicosis caused by AFTN, a maladaptation (17), is more frequently diagnosed in regions with (formerly) scarce iodine supply than in countries with high nutritional iodine intake (18, 19, 20). In preexisting AFTN (21) as well as in preexisting ToxTCa (22), experimental or iatrogenic human exposure to acute increases of iodine supply favored evolution from latent to manifest thyrotoxicosis or exacerbated symptoms of thyrotoxicosis, respectively. In regard to ToxTCa, it is generally thought that only large volumes of malignant tumor tissue might achieve an excessive hormone production, especially in case of large iodine loads. Thus, for AFTN and ToxTCa, the concept of a similarly triggerable molecular pathway of iodine metabolism is plausible, given a comparable differentiation, functionally autonomous volume, and iodine intake.
We suspected a slightly higher frequency of ToxTCa in our patient population during the 1980s, i.e. between 1982 and 1987. Although a transient increase of thyrotoxicoses caused by AFTN has been reported for the same time period (20, 23) after salt iodide concentration had been increased from 7.5 to 15 ppm iodide in Switzerland in 1980 (24, 25, 26), it is uncertain whether a similar impact on ToxTCa frequencies did occur. Case reports of ToxTCa were available in the literature as well as an excellent review by Paul and Sisson (7) but no relevant institutional series. Our aim therefore was to evaluate survival of our own patients with ToxTCa.
Subjects and Methods
Patient files with a history of Tca, with or without thyrotoxicosis, up to 1996 from the Divisions of Nuclear Medicine and Visceral Surgery of the University Hospital, Bern, Switzerland were evaluated. Name, sex, birth date, date of thyroidectomy, date of diagnosis of TCa, histopathologic type, initial TNM classification (27), thyroid function status, and all scintigraphic information were registered from the patient files. Primary histopathologic diagnoses of TCa were adjusted to the 1988 World Health Organization classification. All primary tumors were of thyroidal origin. No case of hyperfunctioning struma ovarii was included (28). Scintigraphic differentiation among cold, warm, or hot nodules and etiologies of thyrotoxicosis have been described previously (18, 19, 20). For patients with follicular or papillary diffTCa and thyrotoxicosis, criteria for a diagnosis of ToxTCa were: 1) thyrotoxicosis and scintigraphic hot thyroid area attributable only to diffTCa (diameter >1 cm, as determined by pathologic examination); and/or 2) thyrotoxicosis caused by radioiodine-retaining metastases with functionally suppressed thyroid/remnants in place; and/or 3) recurring thyrotoxicosis or persisting euthyroidism after total thyroidectomy.
In case the scintigraphically hot nodule was mistaken for a benign AFTN, before histopathologic diagnosis of malignancy was made, the therapeutic activity of 131I was calculated following a standardized European protocol (29), based on an intended dose of 300 Gy to be absorbed by the AFTN volume. If this standardized technique failed to restore euthyroidism, evolution was considered atypical, and a surgical resection was planned. The preferred surgical technique was ipsilateral extracapsular hemithyroidectomy on the side of the toxic nodule, completed by heterolateral subtotal hemithyroidectomy; a neck dissection was performed in case of cervical metastases. If a diffTCa was diagnosed, postoperative metabolic radioiodine therapy following an established cancer protocol was administered (30 30A ). 131I activities aimed at a radiation dose of 500 Gy to the thyroid remnants first, thereafter followed, in case of (suspected) metastases, by standard activities of 7.4 GBq, which were discontinued only after persistent remission had been observed (30 30A ).
Survival of patients with ToxTCa was compared with published survival data of own diffTCa patients (n = 545) with a follicular (50.1%) or papillary (49.9%) subtype (30 30A ). To exclude biases in risk profiles concerning survival of patients with ToxTCa, we matched every ToxTCa patient with three diffTCa patients (the 57 matched diffTCa patients are designated as TCa-M) by age, sex and histopathologic type and corrected statistically for variations in stage distribution M1 against M0 (and T4 against T0–3). Statistical methods included chi-square (2) and Fisher’s tests (frequency comparisons), logistic regression (histopathologic trends), Kaplan-Meier survival curves and Cox regression analysis (survival rates). Anonymity of individual medical data was protected according to Swiss federal regulations on medical ethics. Approval for the study to proceed and monitoring of progress was done by the equivalent of an internal review board.
Results and statistical analysis
In our referral hospital, within 924 TCa patient files, there were 845 diffTCa (408, 40.9%) follicular diffTCa, including 76 oncocytic TCa and 437 (47.3%) papillary diffTCa), 38 anaplastic TCa, 13 medullary Tca, and 28 other histopathologic findings, i.e. 18 uncertain types of diffTCa and, in 10 patients, several types of diffTCa (together 8.5%).
ToxTCa (all Caucasians, Table 1
) was found in 19/924 TCa (2.1%) or 19/845 diffTCa (2.25%) patients and were more frequently follicular (n = 15/19 vs. 4/19 papillary ToxTCa, P = 0.008). Moreover, 2/4 (50%) of the papillary ToxTCa presented with a follicular variant (Table 1). The degree of histopathologic differentiation (data not shown) was not indicative of thyroid hormone status. All tumor diameters were more than 4 cm (range 4–10 cm). ToxTCa were diagnosed most often during the 1980s (Table 1). Average age at diagnosis of ToxTCa patients (65 yr) was much higher than in diffTCa patients (53 yr, P < 0.0005) and slightly higher than in patients with a follicular diffTCa (59 yr, P = 0.025). Male/female sex ratio was 0.73, not different from the diffTCa group (P = 0.135). ToxTCa presented in more advanced stages than diffTCa (T4 in 44% vs. 20.4% of patients, P = 0.045). Follicular and papillary ToxTCa, vs. the 545 diffTCa, presented in T4 stages in 43% vs. 21.5% of patients and in 50% vs. 19.3% of patients, respectively (ns). Frequency of locoregional metastases of ToxTCa was not different from that of all diffTCa (P = 0.285), with N1 stages in 7% vs. 8% of patients with follicular and in 25% vs. 33% of patients with papillary ToxTCa (ns). Frequency of distant metastases of ToxTCa marginally differed from that of all diffTCa (P = 0.052); with M1 stages in 27% against 17% of patients with follicular and in 25% against 7% of patients with papillary ToxTCa (ns). Progressive disease over 8 yr was observed in 88% of ToxTCa patients.
|
|
, patient 2). In patients with ToxTCa, preoperative thyroid uptake of radioiodine was high: 3–82% (median 23%) at 2 h and 5–77% (median 34%) at 24–48 h (Table 1). This includes patient 4 under carbimazole medication and iodine exposure as well as patient 12 with complete functional thyroid suppression because of hyperfunctioning metastases in liver and sacrum (48 h-uptake: 12% and 26%, respectively). In 5 of 19 patients (26%), a range of one to four (median two) therapeutic activity courses of 131I were preoperatively administered to the AFTN that initially had been mistaken for benign (Table 1). Because atypical, recurring thyrotoxicosis required surgical intervention and malignancy was thereby diagnosed histopathologically, a range of one to five (median three) additional 131I therapy courses were administered thereafter (Table 1). Thus, the total number of 131I therapy courses per patient ranged from one to nine (median three). Moreover, palliative percutaneous radiotherapy of metastases was administered in 7 of 19 patients (37%) (Table 1). 131I therapy was administered more frequently to patients with ToxTCa, totalizing higher activities for ToxTCa than for diffTCa, respectively (average 21.8 vs. 15.2 GBq or high activities between 22.2 and 37 GBq in 38.9% vs. 4.9% of patients, P < 0.01, Table 1). In contrast, lesser therapeutic activities of 131I, less than 22.2 GBq were applied in only 55.6% of ToxTCa but in as many as 94.6% of diffTCa patients. Clinical follow-up lasted until death in as much as 90% of ToxTCa patients (n = 19) and in only 25% of diffTCa patients (n = 545). With ToxTCa, median duration of follow-up was 5.9 yr (range 0.8–19.1 yr, 56% over 5 yr, 2 of 19 patients alive and observed for at least 6 yr). With diffTCa, median duration of follow-up was 8.1 yr (up to 25 yr, 65% over 5 yr). Five-year survival rates were lower (56% against 94.5%, hazard ratio 4.8, 95% confidence interval (CI) 2.8–8.1, P = 0.001) in patients with ToxTCa than in those with diffTCa. Ten-year survival rate of 23% with ToxTCa is low, compared with 50% in men with follicular diffTCa and 90% in women with papillary diffTCa (30 30A ).
Because there were differences in age, sex, frequency of follicular subtypes, and presence or absence of distant metastases (and of T4 stages) between patients with a ToxTCa and those with a diffTCa, a bias in the overall risk profile concerning survival was postulated. Therefore, we matched every patient with a ToxTCa with three patients with a diffTCa (TCa-M, of the n = 545 diffTCa group). Practically, for each patient with a ToxTCa, we first selected within the diffTCa group (all papillary and follicular diffTCa, exclusion of oncocytic diffTCa), those patients corresponding by age, sex, and histopathologic type. Thereafter, those diffTCa patients corresponding moreover best by T and M stages were included in the matched TCa-M group. After thus matching for age, sex, and histopathologic type [in both groups, the follicular subtype (79%) predominated], differences in survival were attenuated: The 5- and 10-yr survival rates of the TCa-M group were 74% and 36%, respectively (hazard ratio 2.1, 95% CI 1.13–3.9, P = 0.02, Fig. 1). Correcting thereafter statistically for M1 against M0 stage distribution resulted in a further reduction of the hazard ratio (hazard ratio 1.8, 95% CI 0.93–3.48, P = 0.08). Although a simultaneous mathematical correction for both T4 against T0–3 and M1 against M0 stage distributions was difficult with a data set of only 19 patients, a separate Cox regression analysis showed that the fraction of patients with a T4 stage had a lesser influence on survival rates than the fraction of patients with an M1 stage (hazard ratio of M1 against M0: 2.3). Moreover, a comparison of the ToxTCa and TCa-M groups revealed that the relative risk of a follicular vs. a papillary histopathologic subtype was 1.21.
|
In our overall institutional collective of 924 patients with a TCa, as well as in the published institutional subcollective of 545 patients with a diffTCa (30 30A ), the near-equal, relative distributions of papillary (47.3% and 49.9%, respectively) and follicular (44.2% and 50.1%, respectively) subtypes were typical findings in a region with previously severe iodine deficiency, where during the 1990s mild to moderate iodine deficiency prevailed only in subgroups of the population (25, 31). In contrast, in regions with long-standing high iodine supply, the corresponding distribution may turn up to around eight to one (32). Other Swiss data on distributions of TCa histopathologic subtypes are available. When salt iodization was introduced in Switzerland and iodine supply was still severely deficient in many layers of the population, frequency distributions of TCa had shown a highly significant preponderance of anaplastic (54%) and follicular (26.8%) TCa, but papillary (9.2%) diffTCa was found less frequently (33, 34). In a recent surgical collective from the Zürich region in Switzerland, frequencies of diffTCa presented a slight preponderance of papillary diffTCa (n = 149, 56.4%) vs. follicular diffTCa (n = 115, 43.6%) or, depending on the selection bias, were similar (n = 84, 50.6% for papillary vs. n = 82, 49.4% for follicular diffTCa) (35, 36). In a thorough epidemiologic survey of TCa in the Swiss canton of Vaud from 1974 to 1987, incidence of papillary diffTCa was highest (53%), followed by follicular diffTCa (27%), undifferentiated TCa (5%), and medullary TCa (2%), whereas other morphologies accounted for 13% of the whole series (37). Because there were regionally variable delays in the introduction of iodized salt in Switzerland, prevalences of goiter and incidence of TCa were regionally slightly inhomogeneous, at least up to 1980 (38), albeit being very different from those at the beginning of the 20th century.
ToxTCa is generally noticeable among diffTCa because of the thyrotoxic symptomatology. Thus, ToxTCa was diagnosed more often during the 1980s. The apparent frequency increase by 275%, compared with the 1970s, may have been influenced by ascertainment and methodology. Might the transiently increased frequencies of newly diagnosed ToxTCa presented herein (Table 1) and of published benign AFTN thyrotoxicoses during the 1980s in the same population (20, 23), with a decrease below the baseline thereafter, underlie a similar trigger of preexisting functionally autonomous tissue volume: increased iodine supply (24, 25)? Indeed, urinary iodide concentration in our region had risen from 66 µg/g creatinine in 1978 (n = 77 hospitalized adult patients) to 80 µg/g creatinine in 1991–1992 (n = 30 adult patients with TCa, 24-h urine collections), still indicative of moderate iodine deficiency (24, 25). A slightly increased frequency of benign thyrotoxicoses because of AFTN in Switzerland was considered to be a consequence of rising salt and urinary iodine concentrations since 1980 because thyrotoxic signs, stimulated by increased iodine supply (21), hence diagnosis, occurred prematurely (20, 23). In other parts of the world also, similar events or even real epidemics of thyrotoxicosis have been described, the severity of which might be proportional to the relative increase of iodine supply to iodine-deficient populations (39).
Whether the same causal relation is true for ToxTCa remains hypothetical. Relatively high uptakes of radioiodine at 2 h in our ToxTCa patients point toward an accelerated pathway of iodine metabolism and processing into thyroid hormones. Preferential T3-thyrotoxicosis is often found with functional autonomy (17, 18, 19, 20, 40). Interestingly, T3-hyperthyroidism in our ToxTCa patients was severe (Table 1), as also observed in the review by Paul and Sisson (7), whereas it has been described as mild to severe with benign AFTN (20, 39). Because a sufficient tumor volume and an additional follicular histopathologic component seem to be prerequisites for hormone overproduction, the difference in functional expression between benign AFTN and ToxTCa, given a comparable differentiation, functionally autonomous volume, and iodine intake, might be related to an increased metabolic turnover with ToxTCa, as signaled by high uptakes of 131I. Exceptions to this rule are reported because even small solitary AFTNs of less than 1 cm in diameter are capable of inducing manifest thyrotoxicosis (40).
No case of ToxTCa displayed an oncocytic histopathologic-type TCa; the latter was thus excluded from the matching procedure. In opposition to hyperfunctioning ToxTCa, most cases of oncocytic TCa scarcely concentrate iodine and do not present as ToxTCa. Exceptionally, however, an oncocytic TCa may harbor an activating mutation of the TSH receptor gene (11, 41).
Before matching for age, sex, and histopathologic subtype, patients with ToxTCa had a worse survival prognosis than patients with diffTCa. However, independent characteristics generally linked with higher morbidity and mortality: Higher age at diagnosis, relative preponderance of males, and predominance of follicular type were corrected by the matching procedure. Compared with TCa-M, the not significantly different but possibly higher mortality risk of the 19 ToxTCa patients may be explained by the more frequent occurrence of T4 stages in the latter. Further, albeit M stages were only marginally different, inclusion of this variable into the multivariate Cox model proved decisive in comparing survival rates (hazard ratio 2.3). The existence of metastases in a ToxTCa patient is therefore an important prognostic factor, as is also the case with diffTCa (30 30A ).
Compared with the only published review by Paul and Sisson (7), reporting a 10-yr survival rate of 59% for 43 ToxTCa patients, the 23% rate observed in Bern is low. Our patients were older and more frequently males (65.4 vs. 54.3 yr and sex ratio 0.73 vs. 0.33, respectively). Frequencies of follicular types were comparable. Our patients had distant metastases in only 37% of patients against the 83% mentioned in the review (7). Probably because of relevant therapeutic experiences with ToxTCa, more 131I was administered by us than was reported in the review (average 21.8 GBq in 95% of our patients vs. only 8 GBq in 75% of patients in the review) (7). Our ToxTCa patients were followed up until death or up to a minimum of 6 yr (range 0.8–19.1 yr, median 5.9 yr), whereas in the review follow-up was much shorter (until death or up to a minimum of <0.08 yr, range <0.08–6 yr, median 1 yr). Bernese ToxTCa patients represent a continuous series with strict inclusion criteria vs. a more heterogeneous group in the review, including for instance occult diffTCa.
Conclusion
Clinical and survival data concerning a malignant, rare physiopathologic entity in 19 patients followed up until death or for a period longer than 6 yr in a single institution are presented. Our attention had been attracted by AFTN patients who failed to return to euthyroidism after an 131I therapy course, which was quite unusual. Because cases of early death linked to thyrotoxic crisis were known to occur in ToxTCa patients (7) and our patients with ToxTCa seemed clinically resistant, therefore requiring higher amounts of therapeutic radioiodine and presenting a higher clinical morbidity than those with diffTCa (data not shown), our impression had been that ToxTCa was a clinically aggressive disease with a worse survival prognosis than diffTCa. However, matching procedure and the use of multivariate Cox analyses have shown the biasing influence of age, sex, histopathologic subtype, presentation or not in a T4 stage, and presence or absence of distant metastases. Despite all cited differences between our ToxTCa patients and those mentioned in the review by Paul and Sisson (7) and considering the matching procedures used by us but not in the review, both studies concluded that survival rates for ToxTCa appear not to be much worse than for follicular diffTCa (the latter representing 79% of both the ToxTCa and TCa-M groups).
Higher amounts of 131I applied to our patients with ToxTCa, compared with those with diffTCa, may have positively influenced survival of the former because we clinically observed progressive reduction of thyrotoxic and compressive morbidity and of tumor volume with successive applications of 131I activities. However, this parameter was not formally analyzed. Thus, stringently defined ToxTCa, excluding occult diffTCa, might be considered a relatively well-differentiated, functionally autonomous subtype of follicular diffTCa requiring higher amounts of 131I despite frequently, relatively high tumor uptakes. Confirming findings in the review by Paul and Sisson (7), a ToxTCa should be suspected especially in male patients older than 50 yr presenting with atypical, recurrent, highly symptomatic manifest thyrotoxicosis (data not shown), with a large solitary AFTN greater than 4 cm in diameter and a follicular histopathologic architecture of the tumor. In the present retrospective study, it was not possible to determine the prevalence of TSH receptor and Gs gene mutations. Future multicenter analyses of populations with ToxTCa matched with nontoxic diffTCa would be helpful to describe prevalence of somatic mutations of the TSH receptor and Gs protein genes, natural history, and survival rates more accurately.
Acknowledgments
We thank Manuel Bichsel from the Institute of Social and Preventive Medicine, University of Bern, for performing statistical evaluation; Jane A. Kinser, M.D., and Markus Büchler, M.D., for insight into the patient files; Charles Ruchti, M.D., for diagnostic histopathology; and Prof. Fabio Levi from the Institute of Social and Preventive Medicine, University of Lausanne, for his critical review of the manuscript.
Footnotes
Abbreviations: AFTN, Autonomously functioning thyroid nodule; CI, confidence interval; diffTCa, differentiated thyroid carcinoma; TCa-M, matched patient group with a differentiated thyroid carcinoma; ToxTCa, toxic thyroid carcinoma.
Received July 16, 2001.
Accepted May 9, 2002.
References
genes as a cause of toxic thyroid adenomas. J Clin Endocrinol Metab 82:2695–2701
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
