Papillary microcarcinoma of the thyroid—Prognostic significance of lymph node metastasis and multifocality

Authors


Abstract

BACKGROUND

It is known that patients with papillary microcarcinoma (PMC) of the thyroid gland have a very favorable prognosis. The rising incidence of PMC among papillary thyroid carcinoma (PTC) necessitates the identification of prognostic factors and the formulation of treatment protocols.

METHODS

The authors conducted a retrospective analysis of 203 patients with PMC who were diagnosed on or before 1999 and were treated at the Department of Clinical Oncology, Queen Elizabeth Hospital, Hong Kong.

RESULTS

The cause specific survival, locoregional (LR) failure free survival, and distant metastases failure free survival rates at 10 years were 100%, 92.1%, and 97.1%, respectively. Five patients had lung metastases; 2 patients died of their metastases 12.9 years and 14.8 years after diagnosis, and 3 patients achieved clinical remission after radioiodine (RAI) treatment. Twelve patients had LR recurrences. Patients with LR recurrence were highly salvageable with a combination of surgery, RAI treatment, and external radiotherapy; all but one (who refused treatment) were alive without disease at last follow-up. Multivariate analyses did not reveal any independent prognostic factor for survival. The risk of cervical lymph node (LN) recurrence increased 6.2-fold (P = 0.01) and 5.6-fold (P = 0.02) when LN metastases and multifocal disease were present at diagnosis. RAI ablation reduced the LN recurrence rate to 0.27 (P = 0.04). The presence of LN metastasis increased the rate of distant metastasis 11.2-fold (P = 0.03). Age was not a significant factor in predicting disease recurrence or survival. Subdivision by tumor sizes ≤ 5 mm and > 5 mm did not affect the outcome, but no patient with tumors ≤ 5 mm had mortality related to PMC.

CONCLUSIONS

Despite the overall excellent prognosis for patients with PMC, PMC was associated with a 1.0% disease-related mortality rate, a 5.0% LN recurrence rate, and a 2.5% distant metastasis rate. Therefore, the treatment of patients with PMC should be no different from the treatment of patients with conventional PTC: i.e., complete surgery with consideration for RAI and/or external radiation therapy if poor prognostic factors are present. Cancer 2003;98:31–40. © 2003 American Cancer Society.

DOI 10.1002/cncr.11442

Papillary microcarcinoma of the thyroid (PMC) is defined as papillary carcinoma measuring ≤ 1 cm in greatest dimension according to the World Health Organization classification system for thyroid tumors.1 Before this definition was introduced, confusion in terminology and definitions rendered comparing data a difficult task. These included small papillary carcinomas (e.g., tumors measuring ≤ 1.5 cm2–4) and the descriptive terms occult papillary carcinoma and incidentaloma, which were used to describe incidental finding at autopsy or in thyroidectomy specimens of small papillary carcinoma with no clinical suspicion of malignancy. Some authors included both papillary carcinoma and follicular carcinoma as occult thyroid carcinoma. To avoid confusion, this discussion is restricted to papillary thyroid carcinoma (PTC) because of the distinctive difference in clinical features and outcomes of patients with papillary carcinoma and follicular carcinoma.5 The discussion on thyroidectomy series is limited herein to reports on tumors measuring ≤ 1 cm. The prognoses for patients in thyroidectomy series are excellent.6–11 Some reports revealed zero mortality;7, 9–11 whereas, in other reports, a very low mortality rate of 0.25–0.4%6, 8 was found. PMC can be lethal, because small numbers of patients develop locoregional (LR) recurrences6–8 and distant metastases (DM).8, 11 One study reported a long-term recurrence rate of 6% at 20 years.6

In autopsy studies, the incidence of small PTC varied from 1.0% to 35.6%,12–15 depending on the thoroughness of examination (e.g., the number of sectioning levels) and diagnostic criteria.16 Lymph node (LN) metastases were detected in 3.1–18.2% of patients.14, 15 Marked geographic differences in incidence rates were noted: 1% in Brazil,12 13% in Hong Kong (unpublished data from Queen Elizabeth Hospital), 11.3–28.4% in Japan,4, 13, 17 1.5% in Greece,18 and 35.6% in Finland.15 Based on the high incidence in autopsy studies in Finland, this common incidental finding was viewed as a normal finding, and tumors that measured ≤ 5 mm were considered tumor instead of carcinoma to avoid unnecessary surgeries.15 Most of the autopsy series did not reveal differences in incidence with respect to gender,15, 19 age,15, 20 thyroid size,14 or multifocality.14 Further subdivision of tumors by size (≤ 5 mm vs. > 5 mm) was suggested by Kasai and Sakamoto because of the lower frequency of LN metastasis and extrathyroid extension in the subgroup of patients with smaller tumors.19

Studies in our hospital21 and in Japan22 revealed a temporal trend toward decreasing tumor size of PTC. With advances in ultrasonography and fine-needle aspiration biopsies, especially in mass screening programs, PMC can be diagnosed before surgery.23, 24 Controversies arise because of conflicting reports in management policies. By analyzing 203 patients with PMC who were seen at a single institute, we have attempted to identify the prognostic factors and optimal treatment for this group of patients.

MATERIALS AND METHODS

The records of 997 patients with PTC who were treated at the Department of Clinical Oncology, Queen Elizabeth Hospital from 1960 to 1999 were reviewed retrospectively. Our department is a tertiary referral center for the management of patients with malignant disease. Tumors were classified histologically according to the World Health Organization criteria.1 Of 997 patients, 110 patients with missing information on the size of their primary thyroid tumor and 3 patients with no primary tumor found in resected thyroidectomy specimen were excluded (either PTC found in ectopic sites or metastasis proven by pathologic examination). Among the remaining 884 patients, 203 patients (23%) were diagnosed with papillary microcarcinoma (PMC). In the whole group of 997 patients, a temporal trend toward decreasing mean tumor size was noted: Before 1980, the mean tumor size was 3.25 cm; during 1980–1989, the mean tumor size was 2.45 cm; and, during 1990–1999, the mean tumor size was 2.25 cm (P = 0.001). The percentage of PMC in patients with PTC during these 3 periods increased progressively from 11.9%, to 21.6%, and to 24.5%.

Management Strategy

The management of patients with PTC consisted basically of primary surgery followed by evaluation for radioiodine (RAI) treatment and external radiotherapy, as described previously.25 In general, we prefer bilateral thyroidectomy for patients with PTC, because total or near-total thyroidectomy results in fewer recurrences compared with unilateral surgical procedures. For incidental finding of PTC, we would discuss with patients the options of completion thyroidectomy and RAI ablation. If patients underwent contralateral completion thyroidectomy within 6 months of their initial thyroid surgery, then it was considered part of initial surgery. If patients underwent contralateral surgery after 6 months postdiagnosis, after which tumor in the opposite lobe was confirmed, then it was considered local recurrence. Our criteria for RAI ablation in patients with PTC was tumor size > 1cm, LN metastasis, age older than 40 years, presence of extrathyroid extension, macroscopic postoperative residual disease in the neck, and/or DM. The usual dose for patients with or without distant metastasis was 2.96 gigabecquerel (GBq) (80 mCi) and 5.55 GBq (150 mCi), respectively. External radiotherapy to the thyroid bed and cervical lymphatics was given to patients with macroscopic postoperative LR disease, extensive extrathyroid extension, and LN metastasis. Sometimes, protocol violations happened because of the preferences of individual physicians or patients. After primary treatment, all patients received thyroxin in suppressive or replacement doses. After 1998, the thyroid-stimulating hormone (TSH) level was kept in the normal range for low-risk patients. Serum thyroglobulin monitoring was available after January, 1989. Human recombinant TSH was not available in our hospital. The International Union Against Cancer26/American Joint Committee on Cancer27 TNM staging system was used for the classification of prognoses.

Statistical Analysis

We analyzed the clinical, pathologic, and treatment factors in relation to the following outcome parameters: cause specific survival (CSS), LR failure free survival (LRFFS), and DM failure free survival (DMFFS). LR failure was defined as clinically or radiologically detectable recurrences in the thyroid bed or cervical LNs. DM was clinically or radiologically detectable disease outside the neck. All CSS, LRFFS, and DMFFS curves were generated by the Kaplan–Meier method. Log-rank tests were used to compare prognostic factors. Differences in clinical parameters were analyzed with chi-square tests, t tests, or Fisher exact tests, as appropriate. Relevant variables were entered into multivariate analyses using a Cox regression model.28 SPSS software (version 10.0; SPSS, Inc., Chicago, IL) was used in the data analyses. Significance levels were presented as P values. It was assumed that the observed differences were statistically significant at the P ≤ 0.05 level. Subgroup analyses were performed in relevant circumstances.

RESULTS

The mean follow-up for the 203 patients with PMC was 8.4 ± 5.5 years. Thirty-three patients (16.3%) were lost to follow-up after a mean of 4.9 years. The majority of patients were ethnic Chinese (97%). The CSS, LRFFS, and DMFFS rates at 10 years were 100%, 92.1%, and 97.1%, respectively. Two patients died of disease after 10 years; the 15-year CSS rate was 95%. Most patients presented with Stage I disease (81.8%).

Two patients had DM in the lungs at presentation, and 3 patients developed DM in the lungs (with additional bone metastasis in 1 patient) as recurrent disease at 0.6 years, 9.7 years, and 10.8 years after diagnosis. LR recurrences occurred in 12 patients (in LN in 10 patients and in the thyroid bed in 2 patients). All LR recurrences occurred within 8 years after diagnosis (mean, 3.1 years).

Comparison of Patient Characteristics and Outcome among Patients With and Without PMC

The basic demographic data on patients with and without PMC and are compared in Table 1. At presentation, patients with PMC had a higher female-to-male ratio (6.5 vs. 4.2) and had lower rates of extrathyroid extension (20.7% vs. 31.3%), LN metastasis (24.5% vs. 33.3%), and DM (1% vs. 4%). No patients with PMC were considered inoperable. A lower percentage of patients with PMC received RAI treatment (67.5% vs. 80.9%) and external radiotherapy (3.4% vs. 16.3%). Lower rates of LR recurrence (10.3% vs. 16%) and DM (2.5% vs. 7.6%) were found in the PMC group. A greater proportion of patients had Stage I disease in the PMC group (81.8% vs. 52.4%). CSS, LRFFS, and DMFFS rates were significantly better in the patients with PMC.

Table 1. Patient Characteristics and Treatment for Papillary Microcarcinoma (PMC) Compared with Patients in the Non-PMC Group
Patient characteristicNo. of patients (%)P value
PMCNon-PMC
  1. PMC: papillary microcarcinoma; SD: standard deviation; NS: nonsignificant; UICC: International Union Against Cancer; AJCC: American Joint Committee on Cancer; CSS: cause-specific survival; LRFFS: local-regional failure–free survival; DMFFS: distant metastasis failure-free survival.

Total patients203681
Age (yrs)   
 Mean ± SD46.8 ± 13.445 ± 16.70.17
 Range7.7–77.210.5–91.6
Gender (female:male ratio)6.54.20.048
 Female176 (86.7)549 (80.6)
 Male 27 (13.3)132 (19.4)
Mean size ± SD (cm)0.7 ± 0.282.9 ± 1.7
Multifocal disease   
 No140 (69.0)440 (64.6)NS
 Yes 63 (31.0)213 (31.3)
 Not stated  0 (0.0) 28 (4.1)
Extrathyroidal extension   
 No161 (79.3)346 (50.8)< 0.001
 Yes 42 (20.7)323 (47.4)
 Not stated  0 (0.0) 12 (1.8)
Lymph node metastases   
 No153 (75.4)450 (66.1)0.028
 Yes 50 (24.6)226 (33.2)
 Not stated  0 (0.0)  5 (0.7)
Distant metastasis at presentation  2 (1.0) 27 (4.0)0.036
Type of thyroid surgery   
 Total or near-total thyroidectomy187 (92.1)632 (92.8)0.01
 Lobectomy 16 (7.9) 30 (4.4)
 Biopsy/no surgery  0 (0.0) 19 (2.8)
Lymph node surgery   
 None resected148 (72.9)455 (66.8)0.2
 Excision/sampling 35 (17.2)157 (23.1)
 Neck dissection 20 (9.9) 64 (9.4)
 Not stated  0 (0.0)  5 (0.7)
Radioiodine treatment137 (67.5)551 (80.9)< 0.001
External radiotherapy  7 (3.4)111 (16.3)< 0.001
Locoregional recurrence 12 (10.3)109 (16.0)< 0.001
Distant metastasis  5 (2.5) 52 (7.6)0.008
Lung metastasis during the course of disease (total)  5 (2.5) 39 (5.7)0.06
 At presentation  2 (1.0) 20 (2.9)0.1
 As recurrence  3 (1.5) 19 (2.8)0.4
UICC/AJCC TNM staging   
 Stage I166 (81.8)357 (52.4)< 0.001
 Stage II  0 (0.0) 95 (14.0)
 Stage III 35 (17.2)207 (30.4)
 Stage IV  2 (1.0) 20 (2.9)
 Undetermined  0 (0.0)  2 (0.3)
Ten yr survival (%)   
 CSS100920.002
 LRFFS92.181.90.004
 DMFFS97.191.90.0089

Prognostic Factors

Multivariate analysis did not reveal any prognostic factors for CSS. Table 2 summarizes the LR and LN recurrence analyses. Higher LR recurrence rates were found in patients with LN metastasis at presentation (relative risk [RR], 4.2; 95% confidence interval [95%CI], 1.3–13.6; P = 0.019) and in patients who had no RAI ablation (RR, 0.2; 95%CI, 0.07–0.7; P = 0.01). Multifocal disease (defined as > 1 foci of PTC in the thyroidectomy specimen) was marginally insignificant (P = 0.07). LN recurrence increased in patients with cervical LN metastasis at presentation, multifocal disease, and the absence of RAI ablation. The LN recurrence rate increased 6.2-fold when there was positive LN metastasis at presentation (95%CI, 1.6–24.4; P = 0.01) and increased 5.6-fold when multifocal tumor was found in thyroidectomy specimen (95%CI, 1.3–23.4; P = 0.02). RAI ablation reduced the relative risk of LN recurrence to 0.27 (95%CI, 0.08–0.93; P = 0.04). For DMFFS, LN metastasis was the only prognostic factor, and the presence of LN metastasis increased the relative risk of DM to 11.2 (95%CI, 1.3–100.7; P = 0.03).

Table 2. Locoregional Failure-Free Survival and Lymph Node Recurrence Analyses According to Prognostic Factors
CharacteristicPatients with locoregional recurrencePatients with LN recurrence
No. (%)Univariate P valueMultivariate analysisNo. (%)Univariate P valueMultivariate analysis
RR (95% CI)P valueRR (95% CI)P value
  1. LN: lymph node; RR: relative risk; 95% CI: 95% confidence interval; NS: not significant.

Age        
 ≤ 45 yrs 7/100 (7.0)0.52NS 6/100 (6.0)0.55NS
 ≥ 45 yrs 5/103 (4.9) 4/103 (3.6)
Gender        
 Female11/176 (6.3)0.58NS 9/176 (5.1)0.79NS
 Male 1/27 (3.7) 1/27 (3.7)
Cervical LN metastases        
 No 5/153 (3.3)0.011.00.019 3/153 (2.0)0.00141.00.01
 Yes 7/50 (14.0)4.2 (1.3–13.6) 7/50 (14.0)6.2 (1.6–24.4)
Size of primary thyroid tumor        
 ≤ 0.5 cm 4/70 (5.7)0.97NS 8/133 (7.0)0.37NS
 > 0.5 cm 8/133 (6.0) 2/70 (2.9)
Multifocal disease        
 No 5/140 (3.6)0.051.00.07 3/140 (2.1)0.0051.00.02
 Yes 7/63 (11.1)3.0 (0.9–9.6) 7/63 (11.1)5.6 (1.3–23.4)
Extrathyroidal extension        
 No101/161 (6.2)0.78NS 8/161 (5.0)0.92NS
 Yes 2/42 (4.8) 2/42 (4.8)
Type of thyroid surgery        
 Total/near total thyroidectomy10/187 (5.3)0.21NS10/187 (5.3)0.35NS
 Lobectomy 2/16 (12.5) 0/16 (0.0)
Type of LN surgery        
 Excision 5/35 (14.3)0.048NS 5/35 (14.3)0.0059NS
 Neck dissection 2/20 (10.0) 2/20 (10.0)
 No LN excised 5/148 (3.6) 3/148 (2.0)
Radioiodine treatment        
 No 7/66 (11.4)0.0351.00.01 5/66 (7.6)0.191.00.04
 Yes 5/137 (3.6)0.2 (0.07–0.7) 5/137 (3.6)0.27 (0.08–0.93)
External radiotherapy        
 No10/196 (5.1)0.007NS 8/198 (4.1)0.0012NS
 Yes 2/7 (28.6) 2/7 (28.6)

The significance of LN metastasis at diagnosis is illustrated in Table 3. The increased in LR recurrence rate was accounted for by the increase in LN recurrence rather than local thyroid bed recurrence. Among 50 patients who had LN metastasis at presentation, 14.0% of patients developed LN recurrences and 8.0% of patients had DM, compared with only 2.0% and 0.65%, respectively, in patients without LN metastasis. The incidence of LN metastasis increased with the presence of multifocal disease in primary thyroid tumors (34.9% vs. 20%; P = 0.034). Figure 1 depicts the LN recurrence rate with respect to surgery, initial LN metastasis status, and RAI treatment. RAI ablation in patients with LN negative disease, after undergoing bilateral surgery, resulted in a lower rate of LN recurrence: 7.1% (3 of 42 patients) versus 0% (0 of 95 patients; P = 0.03). In the cohort of 42 patients with N0 disease, the LN recurrence rate was 7.1% after thyroidectomy alone. For those 95 patients who had the same staging but received RAI after surgery, no patient had a recurrence. For patients with N1 disease, the overall LN recurrence rate after RAI ablation was not significantly different (P = 0.6). Age, gender, extrathyroid extension, type of thyroid surgery, and LN surgery were not distinguished as significant factors in any of the analyses, univariate or multivariate, of CSS, LR control, or DM.

Table 3. Lymph Node Metastasis and Outcome
StatusNo. of patients with lymph node metastases at presentation (%)P value
AbsencePresence
Total patients15350
Locoregional recurrence5 (3.3)7 (14.0)0.01
 Local2 (1.3)0 (0.0)1.0
 Lymph node3 (2.0)7 (14.0)0.002
Distant metastasis (all)1 (0.65)4 (8.0)0.006
 At presentation0 (0.0)2 (4.0)0.006
 At recurrence1 (0.7)2 (4.0)0.15
Figure 1.

Lymph node (LN) recurrence with respect to treatment, initial LN node status, and radioiodine (RAI) treatment. The number of patients in each group is shown in parentheses. NS: not significant; +ve: positive; −ve: negative.

Patients Without DM at Presentation

For the 201 patients without DM at presentation, the findings were similar to those for whole group of 203 patients. No single, independent prognostic factor could be found for CSS. The LR recurrence rate was increased 4.5-fold by the presence of LN metastasis at presentation (RR, 4.5; 95%CI, 1.4–15.2; P = 0.014) and was reduced to 20% by RAI ablation (RR, 0.2; 95%CI, 0.06–0.7; P = 0.011). Multivariate analysis showed that the prognostic factors for LN recurrence were the same as the factors for whole group of patients with PTC, i.e., multifocal disease (RR, 5; 95%CI, 1.1–22.3; P = 0.03), LN at presentation (RR, 5.5; 95%CI, 1.3–22.6; P = 0.02), and RAI ablation (RR, 0.2; 95%CI, 0.06–0.9; P = 0.03).

Subdivision of Tumors According to Size

The subdivision of primary tumors according to size (≤ 5 mm vs. > 5 mm) did not have significant impact on patient outcome (Table 4). Seventy patients had tumors that measured ≤ 5 mm in greatest dimension: They presented at a younger mean age (45.3 years vs. 49.6 years for patients with tumors that measured > 5 mm) and had lower frequency of extrathyroid extension (4.3% vs. 29.3%). No difference was found in the frequency of LN metastasis or multifocality. These patients also had a higher rate of lobectomy (14.3% vs. 4.5%) compared with total thyroidectomy (85.7% vs. 95.5%), and they received RAI ablation less frequently (58.6% vs. 72.2%; P = 0.06).

Table 4. Comparison of Clinical Characteristics and Outcome of Patients with Papillary Microcarcinoma, Based on the Size of Primary Tumors
Clinical feature or outcomeNo. of patients (%)P value
Tumor size ≤ 5 mmTumor size > 5 mm
  1. SD: standard deviation; DM: distant metastases; LR: locoregional; CSS: cause-specific survival; LRFFS: local-regional failure-free survival; DMFFS: distant metastasis failure-free survival.

Total patients70133
Mean ± SD   
 Age (yrs)49.6 ± 1345.3 ± 13.40.03
 Tumor size (mm)0.38 ± 0.10.89 ± 0.14< 0.001
 Follow-up (yrs)8.2 ± 5.68.5 ± 5.30.7
Gender   
 Female60 (85.7)116 (87.2)0.8
 Male10 (14.3) 17 (12.8)
Extrathyroidal extension 3 (4.3) 39 (29.3)< 0.001
Lymph node metastasis18 (25.7) 32 (24.1)0.8
Multifocal disease18 (25.7) 45 (33.8)0.27
DM at presentation 1 (1.4)  1 (0.8)1.0
Thyroid surgery   
 Total thyroidectomy60 (85.7)127 (95.5)0.025
 Lobectomy10 (14.3)  6 (4.5)
Lymph node surgery   
 No LN excised54 (77.1) 94 (70.7)0.5
 Excision/sampling 7 (10) 26 (19.5)
 Neck dissection 9 (12.9) 13 (9.8)
Radioactive iodine41 (58.6) 96 (72.2)0.06
External radiotherapy 3 (4.3)  4 (3.0)0.7
Recurrences   
 LR 4 (5.7)  8 (6.0)1.0
  Thyroid bed 2 (2.9)  0 (0.0)0.12
  LN 2 (2.9)  8 (6.0)0.5
 DM 0 (0.0)  3 (2.3)0.55
Status at last follow-up   
 Alive with no disease66 (94.3)129 (97)0.1
 Alive with disease 0 (0.0)  1 (0.8)
  LR 0 (0.0)  1 (0.8)
  DM 0 (0.0)  0 (0.0)
 Died of disease 0 (0.0)  2 (1.5)
 Died of other disease 4 (5.7)  1 (0.8)
Ten yr survival (%)   
 CSS100.0100.00.35
 LRFFS91.992.30.97
 DMFFS100.096.20.47

With regard to the outcome parameters of LR failure and DM, there was no significant difference. Incorporation of the tumor size variable into multivariate analyses did not alter the results in terms of prognostic factor identification.

Nonetheless, none of patients in the group with small tumors died of disease. Two patients decided to undergo completion thyroidectomy at 1.2 years and 7.6 years after lobectomy, after which, contralateral thyroid bed tumor (0.4 cm and 0.3 cm respectively) was confirmed by pathology. Both of those patients belonged to the group with smaller primary tumors. Before their second surgery, they were followed regularly by clinical examination and thyroid bed ultrasonography. One patient had ultrasonographic evidence of a thyroid nodule in the remaining lobe, whereas the other patient had no evidence suspicious of recurrence. These two patients may have considered undergoing completion thyroidectomy years later that revealed multifocal contralateral disease rather than developing a clinical recurrence. However, we still included them with the patients who developed local thyroid bed recurrence. Furthermore, 12 patients had tumors that measured < 1–2 mm, and none of them developed recurrent disease.

Patients who Died of Disease

Two patients died of disease, both from respiratory failure related to lung metastases. The first patient was a lady who was diagnosed with a unifocal, 6 mm PMC at age 74 years. Neither extrathyroid extension nor LN metastasis was present. She underwent subtotal thyroidectomy. Lung and bone metastases were detected at 10.8 years. The patient died 12.9 years after diagnosis. Another patient was a man age 44 years with a 1 cm PMC. He had LN metastasis and multifocal disease at presentation. He underwent total thyroidectomy and unilateral LN dissection. Resection margins were negative. Lung metastasis was found 9.7 years after diagnosis, which led to his death at 14.8 years.

Patients with LR Recurrence

Two patients had local thyroid bed recurrences: They both had primary tumors that measured ≤ 5 mm, as discussed above (see Subdivision of Tumors According to Size). Both patients underwent initial lobectomy and had their thyroid bed tumors discovered after undergoing completion thyroidectomy years later. It is arguable whether this represents natural multifocal disease or an actual recurrence. Both patients were treated successfully with completion thyroidectomy and RAI ablation, and both were alive with no disease at the last follow-up.

Ten patients developed LN recurrences. One patient had lung metastasis at diagnosis. All but one patient remained alive with no disease after treatment, which included surgery alone in two patients, surgery and RAI in five patients, surgery and external radiation therapy (EXT) in one patient, and EXT alone in one patient. One patient refused treatment and was alive with a palpable cervical LN at last follow-up.

Patients with Lung Metastasis

Two of five patients with lung metastasis died of respiratory failure. Both had recurrent, bilateral, macronodular lung metastases. The other 3 patients were alive with no evidence of disease at last follow-up (at 3.7 years, 6.8 years, and 8.9 years). Two of those patients had lung metastases that were detected at diagnosis, and the other patient had lung metastases detected at recurrence. The chest X-ray findings showed an either normal or faint reticulonodular (micronodular) pattern. Two patients were diagnosed with a positive RAI scan, although their chest X-rays were normal. These 3 patients were categorized with clinical complete remissions (negative serum thyroglobulin levels and RAI scans and no radiologic evidence of disease) after treatment with RAI doses of 450 mCi, 380 mCi, and 200 mCi.

DISCUSSION

According to the Hong Kong Cancer Registry data in 1998–1999,29 thyroid carcinoma contributed to 15.7% of the incidence of malignancies among females in the age range 15–34 years, ranking second to breast carcinoma. Our data show that PMC contributes to 20% of all patients with PTC seen in our department in the past decade. PMC, as a specific subgroup of PTC, deserves attention because of its increasing frequency among patients with PTC in clinical practice and the implications for patient management among young adults.

The pattern of practice in treatment varies widely for patients with PTC. The lack of randomized trials, temporal changes in presentation and practice, and conflicting reports on treatment modalities (especially on RAI) all confuse practitioners about what constitutes the optimal treatment for this subgroup of patients. This study showed that PMC was capable of causing mortality (2 of 203 patients; 1%), despite the overall excellent prognosis for these patients. This concurs with studies that included large numbers of patients (178–1628 patients).6, 8, 30 The series that reported zero mortality mostly included smaller numbers of patients (90–120 patients).9–11 The recurrence rates were remarkably lower for patients who had small tumors compared with their counterparts who had larger tumors (> 1 cm). Disease mortality was attributed to lung metastasis, whereas patients with LR recurrence were amenable to treatment with combinations of surgery, RAI, and EXT. The LN recurrence rate was low (5%), comparable to the reported 1.7–6.2% rate in the literature.6, 9, 30

Prognostic Factors

Prognostic factor analyses revealed that LN disease at presentation predicted a higher rate of LR recurrence and DM. Although its significance in predicting survival is disputed, the prognostic value of LN status with regard to recurrent disease has been well documented in studies of patients with PTC25, 31 and PMC.6, 8, 30

The subdivision of tumors according to size did not show a significant difference in outcome, although patients in the group with small tumors did not have disease-related deaths. The 2 patients who had contralateral disease at completion thyroidectomy (> 1 year after primary surgery) may have harbored multifocal disease rather than true local recurrence. The clinical significance, judged by the small number of events, is difficult to ascertain. Furthermore, 12 patients who had tumors that measured < 1–2 mm did not develop any recurrence. This group of patients had tiny tumors that may be considered innocuous.

Among the patients who developed LR recurrence, tumors were as small as 0.3 cm. Most LR recurrences (10 of 12 patients), in fact, were LN recurrences. Whether this is related to the lack of central compartmental lymphadenectomy is a matter of concern.

We found that multifocality was correlated with LN metastasis at presentation, which also was noticed in an autopsy study by Sampson et al.4 We also confirmed that multifocality predicted for LR recurrence, as reported by Baudin et al.7

Older age at presentation is a universally identified poor prognostic factor for patients with PTC.32–38 However, we found that older age did not affect the outcome of patients with PMC adversely, as also observed in the study by Yamashita et al.8 Whether this is related to intrinsic biologic differences or to the overwhelming importance of other factors, like LN status, multifocality, and treatment factors, remains an open question for further studies.

Treatment of Patients with PMC

Total thyroidectomy (or bilateral resection) is the preferred treatment for patients with PTC if their diagnosis of PTC is made before surgery, based on the high rate of multifocal disease (28.3%) and bilateral disease (about 20% at our center). Currently, bilateral surgery is a safe procedure in expert hands. Bilateral lobectomy resulted in a reduction of tumor recurrence rates and cause specific mortality in high-risk patients.39 In low-risk patients, bilateral lobectomy also reduced the recurrence rate.39 Whether completion thyroidectomy is indicated when a lobectomy specimen reveals an incidental PMC should be subjected to individual consideration. The management of patients with PTC varies from completion thyroidectomy (with or without RAI) to observation, sometimes with follow-up imaging studies (i.e., ultrasonography).

Despite the documented effectiveness of RAI for the treatment of patients with differentiated thyroid carcinoma, such as reductions in recurrence rates,25, 32, 35–38 the regression of DM,40–43 and improvements in survival,25, 32, 35, 36, 44 a recently published study of 2444 patients from a single institute refuted these findings.45 The possible reasons for this disagreement may be related to the bulk or volume of tumor left after surgery. If surgery can remove all or most of the tumors, then the effect of RAI in eradicating microscopic foci in thyroid remnants or LN metastasis may be difficult to detect, leading to nonsignificant findings. As a result, one expert in endocrine surgery mentioned that RAI remnant ablation did not improve the already excellent rates of outcome in patients with small PTC.11

For patients with PTC, LN metastasis is a common finding at presentation, varying from ≈ 50% to 80%.46 To date, the recommendation for LN surgery is not unified. In Hong Kong, the current approach for LN surgery is excision for enlarged LNs, with formal neck dissection reserved for patients with extensive LN metastasis. The sensitivity of LN metastasis detection, apart from more extensive surgery, can be increased by special immunohistochemical methods. Micrometastases in cervical LNs were detected in 27% of patients who had PMC with pN0 by immunohistochemical methods.47 It is not surprising that central compartmental LN dissection was advocated as part of surgery in guidelines for the treatment of patients with PTC.47–49 However, dissection should be performed very carefully to avoid damaging the recurrent laryngeal nerves and the parathyroid glands. Our data from Figure 1 revealed that RAI decreased the LN recurrence rate from 7.1% to 0.0% in patients with who were negative for LN metastasis at presentation. It is tempting to speculate that RAI may eradicate microscopic metastasis in LNs. If a central compartmental lymphadenectomy is not performed routinely, which is the current scenario in our locality, then the expected incidence of microscopic LN metastasis after thyroidectomy would be high. The clinical significance of these micrometastases is difficult to predict. With this caveat in mind, it may be a good policy to support the use of RAI ablation in regions where lymphadenectomy is not a routine practice. Undoubtedly, this speculation requires proof in large-scale, systematic studies. To date, the contemporary literature on LN surgery and its implications for recurrences do not provide sufficient evidence to guide the best surgical approach.

Role of RAI

In the current study, RAI reduced the LN recurrence rate, especially in patients with pT1N0 disease. It is possible, as discussed above, that this reduction was related to the bulk of disease. In this group of patients who did not undergo formal LN dissection, the bulk of micrometastases may have been small. In this situation, the effect of RAI in eradicating the disease may be demonstrated more easily.

For patients who were diagnosed with DM, the response to RAI was good. Among the five patients who had DM, three patients achieved a complete clinical remission with negative serum thyroglobulin levels at last follow-up. It is interesting to note that two patients were diagnosed with DM during the initial postoperative period by posttherapy scans, when DM was not noticeable or was present faintly as reticulonodular shadows on chest X-rays. Posttherapy scans show greater sensitivity for detecting DM compared with diagnostic scans, because the dose of RAI is higher. Without these scans, diagnosis might have been delayed for these patients. The chance of a cure is greatest when DM was are RAI avid and while the volume of disease is low (i.e., radiolologically undetectable by chest X-rays and computed tomography scans). With the observation of two fatalities from lung metastases as macronodular recurrences and the success with treating early lung metastasis, it is logical to suggest that early postoperative RAI can detect early DM and successfully treat or even cure the DM.

CONCLUSIONS

The treatment of patients with PMC is controversial. We believe that tumor size should not be the only factor considered. Although total or near-total thyroidectomy was considered the surgical treatment of choice in some centers,11 observation was adopted as the protocol at the other extreme.30 We are looking forward to the long-term results from Sugitani et al., who observed the clinical course of patients with asymptomatic PMC without proceeding to any treatment.50 Their new protocol of conservative treatment started in 1995. Twenty-one patients were followed without disease progression until 1998.30 However, the surveillance policy for patients with PMC is difficult to implement, because our patients have trouble accepting no treatment for malignant disease in which recurrence and mortality definitely are observed, which was as long as > 10 years in this study.

In conclusion, we advise bilateral thyroidectomy if patients who have PTC are diagnosed before surgery, irrespective of tumor size. If poor prognostic factors, such as LN metastasis or multifocality, are present, then RAI may be a good option to reduce the risk of recurrence. Furthermore, the use of RAI during the postoperative period has a dual benefit of diagnosing early DM and treating these micrometastases. Whether the early application of RAI as ablation may improve patient outcome is a matter of debate, and further studies are encouraged.

Acknowledgements

The authors thank Mr. Oscar Mang of the Hong Kong Cancer Registry for his generous support in providing mortality data for analysis.

Ancillary