Patients with thyroid microcarcinoma (TMC) have favorable long term prognoses. However, recurrences in the neck and distant metastases have been reported. The authors investigated independent factors associated with recurrence in an effort to define therapeutic guidelines.
Two hundred eighty-one patients (207 females, 74 males; mean age, 41.9 years) with a differentiated thyroid carcinoma ≤1 cm in greatest dimension (mean size ± standard deviation, 5.9 ± 3.3 mm) were analyzed. The median follow-up time was 7.3 years.
TMC diagnosis was incidental in 189 patients, and metastases were the first manifestation of the disease in the other 92 patients. Therapy included near-total thyroidectomy for 195 patients, lymph node dissection for 195, and therapeutic administration of radioiodine for 124. Eleven recurrences (3.9%) were observed 4.3 ± 2.7 years (mean ± standard deviation) after initial treatment: all had locoregional recurrence (4 in the thyroid bed and 7 in the lymph nodes), and in one of these the local recurrence was associated with lung metastases. Multivariate analysis showed that two parameters significantly influenced TMC recurrence, namely, the number of histologic foci (P < 0.002) and the extent of initial thyroid surgery (P < 0.01). Only 3.3% of patients with unifocal TMC treated with loboisthmusectomy had tumor recurrence.
Thyroid microcarcinoma (TMC) is defined by the World Health Organization as carcinoma 1.0 cm or less in greatest dimension.1 TMC is being found in an increasing proportion of treated differentiated thyroid carcinomas (DTC) due to the more frequent use and improvement of ultrasonography, fine-needle aspiration biopsy, and pathologic procedures.
In DTC patients, multivariate analyses have demonstrated the prognostic importance of tumor size:2-5 no increased mortality has been demonstrated for tumors with a greatest dimension of less than 1.5 cm.5 The high incidence of TMC found in autopsy studies (up to 35%) suggests that most of them have benign behavior.6-14 Conservative treatment, i.e., unilateral lobectomy, has therefore been advocated for patients with these tumors.15 However, locoregional recurrence has been reported in 0-11% of TMC patients,5, 12-18 and rare cases of distant metastases have also been reported.19 In a recent study of 535 cases of papillary TMC, two risk factors of recurrence were identified, namely, lymph node metastases and extent of thyroid surgery. These authors concluded that TMC "has an excellent prognosis if managed initially by bilateral lobar resection."16
The aim of the current study was to individualize factors associated with recurrence in a series of 281 patients with TMC treated at the Institut Gustave-Roussy (IGR) in Villejuif, France, in order to define therapeutic guidelines.
PATIENTS AND METHODS
Two hundred eighty-one patients with DTC 1 cm or less in greatest dimension were consecutively treated and followed up at IGR between 1962 and 1995. All histologic diagnoses were made by one pathologist (B.C.) according to the recommendations of the World Health Organization. Pathologic examination was guided by macroscopic examination; furthermore, normal thyroid areas were routinely sectioned every 3-5 mm. The mean age at initial surgery was 41.9 years (standard deviation [SD], 13.8 years; range, 6-75 years). There were 207 females and 74 males. The mean duration of follow-up was 7.3 years (SD, 6.2 years; range, 0.6-33.7 years). The latest IGR outpatient visits occurred from 1995 for 64% of patients and from 1991 for 90% of patients. At the end of the study, 11 patients had died of causes other than DTC.
When a lobectomy had been performed and TMC was diagnosed after a complete histologic examination, thyroidectomy and/or neck lymph node dissection and radioiodine therapy were considered for each case, with prognostic factors taken into account.20 The lack of a standardized approach to treating TMC patients during these years permitted the study of the impact of various therapeutic modalities. All patients received levothyroxine (LT4) therapy after initial surgery. In all patients, a clinical evaluation with thyroid-stimulating hormone (TSH) and thyroglobulin (Tg) measurements (Dynotest; Henning, Berlin, Germany) was performed 3 months after the initial treatment and then on a yearly basis, as previously described.21 An iodine-131 (131I) total body scan was performed 4 days after the ablative dose of 3.7 GBq 131I (100 mCi); it was performed with a 74 MBq 131I (2 mCi) dose 6 months after the initial treatment for patients treated by total thyroidectomy only. A neck ultrasonography was performed according to clinical and Tg results. Any evidence of tumor arising 6 months or more after initial surgery was considered a recurrence.
Univariate analyses between covariates and risk of relapse were performed using Fisher's exact test of independence. The multivariate prognostic value of covariates with regard to risk of relapse was assessed using exact logistic regression. Logistic models were built in a forward sequential manner, entering at each step the most significantly associated covariate until a covariate was no longer associated with the risk of relapse. Binary variables were as follows: history of external radiotherapy, gender, mode of diagnosis (incidental or not), neck lymph node metastases, histologic type, extrathyroidal and extranodal extent, number of TMC foci (multifocality was defined as more than one focus of TMC and compared with unifocality), unilaterality or bilaterality, vascular invasion, extent of thyroid surgery (total thyroidectomy was compared with loboisthmusectomy or isthmusectomy), lymph node neck dissection, and radioiodine ablation. Age at diagnosis and tumor size were recorded using cutoff values of 40 years and 5 mm, respectively.
Tables and univariate results were computed using SAS software (SAS Institute), whereas LogXact software (Cytel) was used to perform the exact logistic regression. All tests were two-tailed, and P values <0.05 were considered statistically significant.
Clinical Characteristics and Modes of Diagnosis
For 189 patients, TMC was classified as "incidental" because it was found at routine histologic examination after thyroid surgery was performed on 185 patients for benign adenoma, 3 patients for Graves' disease, and 1 patient for laryngeal carcinoma. For 92 patients, TMC was classified as "not incidental": thyroid surgery was performed on 89 patients because lymph node metastases were clinically found in the neck (16 of these 89 also had a palpable thyroid nodule) and 3 patients because of distant metastases (2 in lungs and 1 in bone). In these patients, the thyroid origin of the metastases was ascertained before thyroid surgery by histology and immunohistochemistry using anti-Tg antibodies on a surgical biopsy of a lymph node or a distant metastasis.
Significant differences for the following parameters were found between these 2 groups of patients (Table 1): patients with "not incidental" TMC were significantly younger, the gender ratio was nearly equal to 1, and they had a higher frequency of multicentricity, bilaterality, extrathyroid involvement, and lymph node metastases. Finally, radioiodine was more frequently administered for ablation.
Table 1. Characteristics of the 281 Patients with TMC at Presentation
Surgical treatment consisted of a total thyroidectomy for 195 patients, a loboisthmusectomy for 83, and an isthmusectomy for 3. A homolateral dissection of the jugulocarotid chain and of the paratracheal groove was performed on 195 patients; it was bilateral for 63 of them. Radioidine therapy was administered within 6 weeks after thyroid surgery to 124 patients; external radiotherapy to the neck and mediastinum was given to 10 patients (treated before 1970). The postoperative evaluation revealed lung metastases in three patients and brain metastases in two. These five patients were considered to have initial distant metastases and were excluded from the analysis of recurrences. Clinical and histologic features of these five patients and of the other three with initial distant metastases are given in Table 2.
Table 2. Clinical and Pathologic Characteristics of the 8 TMC Patients with Initial Distant Metastases
The median tumor size was 5.9 mm (SD, 3.3; range, 1-10 mm) (Fig. 1 (6K)). A papillary carcinoma was found in 247 patients and a follicular carcinoma in 34. The tumor was multifocal in 112 patients (40%). One hundred ninety-five patients, 86 with multifocal TMC and 109 with unifocal TMC, underwent a total thyroidectomy; 46 (53%) of the 86 patients with multifocal TMC had bilateral disease. In order to evaluate the risk of underestimating bilateral disease in patients with a single focus of TMC after lobectomy, we looked at the distribution of TMC in patients with multifocal, bilateral TMC treated with total thyroidectomy. Among these 46 patients, 19 had multifocal TMC in both lobes and 27 had a single focus in one lobe associated with 1 or more neoplastic foci in the isthmus or in the opposite lobe. Thus, among the 136 patients who underwent a total thyroidectomy and had a single focus in 1 lobe, 27 (19.8%) had bilateral disease. An extracapsular extension into the surrounding structures was found in 42 patients (14.9%) and vascular invasion in only 10 patients.
Homolateral lymph node metastases were present in 121 (43%) of the TMC patients, bilateral in 20 (7%). Considering the patients who underwent a lymph node dissection, 62% were positive, 38% were in the "incidental" group, and 93% were in the "not incidental" group. Homolateral jugulocarotid and paratracheal metastatic lymph nodes were found in 56% of patients (103 of 182) and 48% of patients (77 of 159), respectively. Bilateral jugulocarotid and paratracheal metastatic lymph nodes were found in 41% (20 of 49) and 27% (12 of 45), respectively. When patients with unifocal or multifocal TMC were compared, lymph node metastases were found in 52% and 75% of patients, respectively.
Postoperative Tumor Recurrence and Survival
Eleven patients (3.9%) had recurrence after a mean follow-up of 4.3 years (SD, 2.7 years; range, 1.4-9.7 years) after initial surgery. Two patients had recurrence within 2 years, 5 between 2 and 5 years, and 4 between 5 and 10 years after initial surgery. Histologic confirmation of the recurrence was obtained for 9 of these 11 patients; in the other 2 cases, the diagnosis was confirmed by the presence of uptake of radioiodine outside the thyroid bed on posttherapy scan. Ten patients had an isolated locoregional recurrence, four in the thyroid bed and six in neck lymph nodes; and one patient had both a neck lymph node recurrence and lung metastases. Recurrence in the thyroid bed occurred in three patients in the opposite lobe after a unilateral loboisthmusectomy and in one patient in the thyroid remnant after a total thyroidectomy. No extrathyroidal involvement was noticed, either initially or at the time of reoperation for recurrence, in these four patients. Neck lymph node recurrence was discovered in five patients who had previously undergone a lymph node dissection, which was positive in four; the recurrence was homolateral in one case, contralateral in three, and bilateral in one. One patient without initial lymph node dissection experienced an homolateral neck lymph node recurrence.
Eight of the 10 patients with locoregional relapse are currently free of disease with an undetectable Tg level following LT4 withdrawal, after a mean follow-up of 150 months (range, 6-345 months) after discovery of the recurrence. One patient with Marfan's syndrome had palpable neck lymph nodes and died shortly after their discovery from aneurysm. In 1 patient, Tg was undetectable during LT4 therapy but increased to 6 ng/mL following LT4 withdrawal, 71 months after treatment of the recurrence; however, the posttherapeutic scan (3.7 GBq 131I) did not show any significant ectopic uptake.
One patient, whose papillary carcinoma measuring 1 cm in greatest dimension was revealed by palpable neck lymph nodes, experienced both a bilateral neck lymph node recurrence and lung metastases 64 months after initial surgery. This patient is still receiving treatment.
Prognostic Factors Associated with Recurrence
Because all patients who relapsed had a locoregional recurrence, and because of the small number of recurrences in the thyroid or in the lymph nodes, they were analyzed together for prognostic factors of recurrence. The eight patients whose distant metastases either revealed TMC (three patients) or were discovered at the first postoperative work-up (five patients) were excluded from this analysis, but details are given in Table 2.
In the multivariate analysis, only two parameters were found to influence TMC recurrence significantly (Table 3): the number of TMC foci at pathologic examination (P < 0.002) and the extent of initial thyroid surgery (P < 0.01). The recurrence rate for patients with unifocal TMC was 1.2%, compared with 8.6% for patients with multifocal TMC. After total thyroidectomy, the recurrence rate was 2.3% compared with 8.2% after loboisthmusectomy or isthmusectomy. When thyroid recurrences were studied as a function of unicentricity or multicentricity and extent of thyroid surgery (Table 4), there were only 2 relapses (3.3%) among the 60 patients with a unifocal TMC who underwent a loboisthmusectomy, one in the contralateral lobe and the other one bilateral in neck lymph nodes. For patients with multifocal TMC, the recurrence rates were 20% and 5% after loboisthmusectomy and total thyroidectomy, respectively. When we sought to determine the prevalence of the main therapeutic procedures, we found, for unifocal and multifocal TMC, respectively, a prevalence of total thyroidectomy of 64% and 76% (P = NS), a prevalence of lymph node dissection of 59% and 83% (P < 0.001), and a prevalence of radioiodine administration of 38% and 54% (P < 0.007).
Table 3. Characteristics of the 11 Patients Who Experienced Tumor Recurrencea
Three circumstances may lead to the discovery of a TMC: surgery for thyroid micronodules with suspect fine-needle aspiration cytology, surgery for benign thyroid disease, and discovery of neck lymph node or distant thyroid metastases for which the thyroid origin has been shown by positive antithyroglobulin immunohistochemistry. These different modes of diagnosis may reflect different populations regarding TMC behavior. Many differences between "incidental" and "not incidental" TMC patients have been found, as already reported.13 We analyzed the mode of presentation as an independent variable in our study. The mode of presentation was not related to the size of TMC. The incidence of lymph node metastases at neck dissection was 38-93% in the "incidental" and "not incidental" groups, respectively, in accordance with previous clinical studies in which this rate ranged from 31% to 79%.12, 13, 15-18 However, in these studies, a lymph node dissection was performed only if obviously involved lymph nodes were present, and they could not be directly compared with our results. This high rate of neck lymph node positivity in TMC found by routine neck dissection, even in the "incidental" group, may explain the predominance of neck lymph node recurrence found by others.13, 16 Regarding the referral pattern at our institution, the TMC under study should be considered aggressive TMC, as shown by the one-third of TMC diagnoses after the discovery of lymph node involvement and the 42 tumors with extracapsular involvement. These features explain what can be regarded as an aggressive therapeutic approach.
Only a few studies have reported long term follow-up of TMC patients: the majority of TMC recurrence was locoregional, in the thyroid bed and neck lymph nodes, and occurred in 0-11% of patients, mostly during the first 10 years of follow-up.5, 12-18 Thus, the mean TMC follow-up of 7.3 years in our study allowed the discovery of the majority of recurrences, if not all of them. Most recurrence was cured after further neck surgery. Our data are in accordance with previous series, and we conclude-as did others-that the recurrence rate is low and does not affect survival. Distant metastases, as found in this study, were unusual but have already been described5, 12-14, 16, 19; finally, death related to TMC has been reported in very few cases.12-14, 16
Hay et al. performed a multivariate analysis of 535 TMC patients with a mean follow-up of 17.5 years (of whom 27 had recurrence) and individualized 2 risk factors for locoregional recurrence, namely, initial lymph node metastases and extent of initial thyroid surgery.16 These authors concluded that a bilateral lobar resection seemed to be the treatment of choice for these patients. Our multivariate analysis individualized two risk factors, namely, multifocality and extent of initial thyroid surgery. However, because the number of recurrences in the thyroid bed and neck lymph nodes was small, the statistical analysis took all recurrences into account. The current study suggests that a more aggressive therapeutic approach should be taken in cases of multifocality. Multifocality is a well-known feature of papillary carcinoma.22, 23 Its rate is highly dependent on the thoroughness of pathologic examination. Many authors have justified total thyroidectomy for patients with papillary carcinoma by its high rate of multifocality and claimed that this should reduce the rate of thyroid recurrence.24 However, multifocality has not been demonstrated as a risk factor for recurrence by retrospective multivariate analyses of large populations of DTC patients.2-5 It is noteworthy that, in our study, microscopic involvement of the contralateral lobe could not be ascertained when a loboisthmusectomy had been performed. However, the risk of recurrence for the group of patients with a single TMC was low, even after loboistmusectomy; thus, we believe that, even if multifocality is not ruled out in this population, its individualization is relevant to therapeutic guidelines. This study did not confirm the prognostic influence of positive neck lymph nodes on the recurrence rate. It should be mentioned that the influence of lymph node metastases on DTC survival is still being debated.25 Neither clinical nor histologic lymph node involvement were prognostic factors for recurrence. However, multicentricity and neck lymph node metastases are strongly interrelated, as suggested by many studies7, 23 and confirmed in the current study. Differences in patient presentation, therapeutic approach, and thoroughness of pathologic examination may modify the incidence rate of both parameters and thereby influence statistical results. Although lymph node involvement had no prognostic impact on the recurrence rate in our study, we still recommend a dissection of the central compartment of the neck in patients with multifocal tumor, for the following reasons: 1) neck lymph nodes are the main site of locoregional recurrences in TMC patients, and 2) involvement of neck lymph nodes is strongly related to multifocality.
The extent of thyroid surgery was also found to have an independent influence on the recurrence rate. It is generally agreed that TMC should not be overtreated.2-5, 12-16 Furthermore, the low rate of recurrence in patients with unifocal TMC should be emphasized. Even after loboisthmusectomy, there were only 2 recurrences among the 60 patients (3.3%), including only 1 recurrence in the remaining lobe. Considering this low incidence of recurrence in patients with unifocal TMC and the potential morbidity of total thyroidectomy, we suggest that conservative thyroid surgery is suitable for patients with unifocal TMC, especially when the diagnosis of TMC is based on final histologic examination. These clinical data should be considered in view of the 19.8% incidence of bilaterality in patients with a unifocal TMC in 1 lobe who were treated with total thyroidectomy. Only a small minority of these foci will recur, as has already been suggested for DTC patients.26 On the other hand, multifocal DTC are bilateral in more than half of cases, and the influence of the total thyroidectomy is obvious: the rate of recurrence decreases from 20% after lobectomy to 5% after total thyroidectomy. We thus recommend a total or subtotal thyroidectomy leaving no more than 2-3 g of thyroid tissue in patients with multifocal TMC at presentation. It should be observed that no significant difference was found regarding the incidence of total thyroidectomy in TMC patients with unifocal or multifocal TMC. Radioiodine administration may be indicated for these patients after total thyroidectomy, but no significant decrease of recurrence was observed in TMC patients treated with radioiodine. It is noteworthy that no locoregional recurrence occurred in patients with initial extrathyroidal involvement.
Surprisingly, the mode of diagnosis was not found to influence the recurrence rate: five and six patients in the "incidental" and "not incidental" group had recurrence, respectively. However, the low incidence of recurrence in this study precludes any conclusion regarding factors not found to be associated with recurrence.
We conclude that TMC is a tumor associated with a good prognosis. Multifocality should be the main factor to consider in selecting the type of thyroid surgery. A loboisthmusectomy can be considered sufficient for patients with unifocal TMC. On the other hand, a total thyroidectomy with dissection of the central compartment of the neck is recommended for patients with multifocal TMC; administration of radioiodine may also be indicated for patients with multifocal TMC.
The authors are indebted to Elisabeth de la Genardière, Catherine Martin, and Ingrid Kuchental for secretarial assistance.