Percentage of patients treated between 1965 and 1990, by tumor size.
Percentage of patients treated between 1965 and 1990, by age group.
Disease-specific survival curves. Group 1 patients underwent surgery for papillary thyroid cancer during the period 1965-1973; group 2, during the period 1974-1982; and group 3, during the period 1983-1990. For group 1 vs group 2 and for group 1 vs group 3, P =.03 and .04, respectively.
Disease-specific survival curves for patients 50 years or older with tumor 25 mm or larger in the maximum diameter. For a description of the groups, see the legend for Figure 1. For group 1 vs group 2 and for group 1 vs group 3, P=.02 and .01, respectively.
Yamashita H, Noguchi S, Yamashita H, Murakami N, Watanabe S, Uchino S, Kawamoto H, Toda M, Nakayama I. Changing Trends and Prognoses for Patients With Papillary Thyroid Cancer. Arch Surg. 1998;133(10):1058–1065. doi:10.1001/archsurg.133.10.1058
To analyze differences in the demographic backgrounds, and in treatments, prognosis, and risk factors of patients with papillary thyroid cancer operated on from 1965 to 1990, by dividing them into 3 chronological groups.
Retrospective cohort study of 2423 patients with papillary thyroid cancer (tumor size, ≥10 mm) who underwent curative surgery at the Noguchi Thyroid Clinic, Oita, Japan.
A center for the treatment of thyroid disease, at which about 1400 thyroid operations are performed per year.
There were 596 patients treated during from January 1, 1965, to December 31, 1973; 964 patients treated from January 1, 1974, to December 31, 1982; and 959 patients treated from January 1, 1983, to December 31, 1990.
Of the 2519 patients treated, 96 were excluded from the study because they had undergone noncurative surgery. Therefore, the analyses are based on data for 2423 patients who underwent curative surgery. Three groups were defined as follows: group 1, underwent surgery during the period 1965-1973 (n=577); group 2, underwent surgery during the period 1974-1982 (n=924); and group 3, underwent surgery during the period 1983-1990 (n=922). The mean age of the patients in group 1 was 42.4 years, in group 2, 45.0 years, and in group 3, 47.8 years. The mean tumor size was 30.4 mm, 26.5 mm, and 24.6 mm, respectively, for groups 1, 2, and 3. The 10- and 20-year disease-specific survival rates were significantly improved from group 1 (95.5% and 90.3%, respectively) to group 2 (97.8% and 93.9%, respectively), and the 10-year rate was significantly improved for group 3 (98.2%). In the multivariate analysis, age, sex, tumor size, and gross nodal metastasis were significant predictors of survival for group 1; however, only age and gross nodal metastasis were significant for group 3.
Over time, papillary thyroid cancer has become diagnosed at an earlier stage, but the age of the patients at diagnosis is older. The disease-specific survival rate was significantly improved, mainly owing to earlier treatment, and the change in the risk factor profile for cancer mortality may be due to the changes in the demographic backgrounds and diagnostic and therapeutic modalities. These considerations derived from risk factor analysis should be considered for treating the patient and for the prediction of patient survival.
AS A RESULT of the indolent nature of differentiated thyroid cancer, long-term follow-up of patients is necessary to evaluate the effectiveness of treatment modalities. However, the demographic backgrounds of the patients, the behavior of thyroid cancer, and the diagnostic and therapeutic modalities change over long periods, suggesting that misleading information could be generated by the analysis of long-term follow-up data. We must be aware of the kinds of changes that occur and whether these changes affect the outcome for patients.
The age at operation and size of the tumor were reported to be significant risk factors for patients with differentiated thyroid cancer in most studies.1- 7 Owing to the recent increases in life expectancy and the low birth rate, the proportion of older people has increased, while that of younger people has decreased in most industrialized countries. Fine needle aspiration cytologic examination has revolutionized the evaluation of thyroid nodules during the last 2 decades,8,9 and, therefore, early diagnosis and treatment of thyroid cancer can be expected. There have been numerous studies of the treatment of and prognosis for patients with differentiated thyroid cancer1- 7,10; nevertheless, little information is available on the changing trends for patients with thyroid cancer and on the effects of the changing trends on prognosis.11- 14 We therefore analyzed differences in the demographic backgrounds, treatments, prognosis, and risk factors for patients with papillary thyroid cancer treated during 3 chronological periods.
Between January 1, 1965, and December 31, 1990, 3426 patients with papillary thyroid cancer sought treatment at our hospital. Among them, 861 patients with a primary tumor less than 10 mm in the maximum diameter and 46 who had another histologic type of thyroid cancer were excluded from the study. Therefore, we reviewed the medical records of 2519 patients. All patients were Japanese.
The patients were divided into groups according to the period during which they were first treated: January 1, 1965, to December 31, 1973, 596 patients; January 1, 1974, to December 31, 1982, 964 patients; and January 1, 1983, to December 31, 1990, 959 patients. Although surgical procedures have changed gradually, the present procedures were established around 1965. The thallium scintigraphic scan for preoperative examination and follow-up and routine fine needle aspiration cytologic examination of thyroid nodules were introduced in 1983 in our institution. The measurement of serum thyroglobulin was started in 1981 in our institution and became a routine test in 1983.
The patients were followed up until August 1996. The mean and median follow-up periods for the patients known to be alive were 14.6 years and 13.8 years, respectively. (For group 1 patients, mean and median follow-up periods were 25.0 and 24.6 years, respectively; group 2, 16.3 and 16.2 years, respectively; and group 3, 8.2 and 8.0 years, respectively.) Follow-up data were obtained by periodic correspondence with the patients, referring physicians, family members, and/or hospitals or from autopsy and municipal records. Six patients (0.2%) could not be located within 5 years of surgery. The first recurrence was categorized as that in the remnant thyroid, local recurrence (tumor in the neck area), or distant metastasis.
Statistical analyses were performed by using 1-way analysis of variance followed by the Tukey-Kramer multiple comparison procedure. Frequency data were compared using the Fisher exact test (2-tailed). The postoperative survival rate was estimated by using the Kaplan-Meier method. Differences in the disease-specific survival and recurrence-free survival curves were estimated by using the Wilcoxon test. Univariate and multivariate analyses were performed by using the Cox proportional hazards regression model15 with data analysis software (SAS JMP, version 3.2R SAS Institute, Cary, NC) for the Macintosh computer (Apple Computer Inc, Cupertino, Calif). The relative importance of variables was expressed as the risk ratio and tested by calculating P values. A distinction was made between death due to thyroid malignancy and other causes, based on the death certificate, the medical records, or both.
Of the 2519 patients, 2278 were female and 241 were male (female-male ratio, 9.5:1). Their ages ranged from 6 to 90 years, with mean age of 45.9 years. The mean patient ages were as follows: 1965-1973, 43.0 years; 1974-1982, 45.3 years; and 1983-1990, 48.3 years. Of the 2519 patients whose medical records were reviewed, 96 underwent noncurative surgery because of locally advanced disease, distant metastatic lesions, or both. The distribution was as follows: 1965-1973, 19 (3.2%) of 596 patients; 1974-1982, 40 (4.1%) of 964 patients; and 1983-1990, 37 (3.9%) of 959 patients. Data for these 96 patients were excluded from the analyses reported in the present article; therefore, the analyses are based on data for 2423 patients who underwent curative surgery (Table 1). The 3 groups were defined as follows: group 1, underwent surgery during the period 1965-1973 (n=577); group 2, underwent surgery during the period 1974-1982 (n=924); and group 3, underwent surgery during the period 1983-1990 (n=922). The mean age at diagnosis was significantly increased throughout the periods represented by the groups. The percentage of patients aged 50 years or older was 30.2%, 37.1%, and 49.0% in groups 1, 2, and 3, respectively; the difference was statistically significant (P<.001). The mean of the maximum diameter of the papillary cancer showed a significant decrease throughout the periods represented by the groups (Table 1). The percentages of tumors smaller than 20 mm were 18.0%, 31.0%, and 40.2% for groups 1, 2, and 3, respectively. The percentages of patients by tumor size and age (Figure 1 and Figure 2) show that papillary thyroid cancer is being diagnosed when the tumor is smaller, but that patients are older when first given the diagnosis.
The frequency of the preoperative diagnosis of the papillary cancer was significantly improved from 60.1% to 73.7% and then to 79.6% for groups 1, 2, and 3, respectively (Table 1). The decreased percentage of patients with gross extracapsular invasion was significant between groups 1 and 2 (P<.001). There was no distinct difference in the percentage of gross or microscopic nodal metastasis.
Among the 2423 patients, 628 (25.9%) were given a diagnosis of a benign thyroid disease, before or at pathological examination (Table 2). Benign thyroid nodules (adenomatous goiter in 10.7% and follicular adenoma in 7.2%) occurred in 17.9%, and autoimmune thyroid disease (chronic thyroiditis in 6.7% and Graves disease in 3.0%) occurred in 9.6%.
Table 3 gives the operative procedures used for the treatment of papillary carcinoma for each group. Total thyroidectomy was used in only 2.6% of the patients. Whereas the frequency of subtotal thyroidectomy increased, that of partial lobectomy progressively decreased. Regarding the extent of thyroidectomy, the percentages of patients who underwent thyroid lobectomy or more extensive excision were 74.5%, 88.5%, and 95.4% in groups 1, 2, and 3, respectively.
Cervical lymph nodes were excised in 2040 patients (Table 4). The percentage of patients who underwent modified radical neck dissection was highest in group 2.
The data for the postoperative complications after the thyroid surgery are summarized in Table 5. No operative deaths occurred. Until 1981, vocal cord motility was examined only if the patient complained of hoarseness; thereafter it was examined in all patients before and after surgery. None of the patients had permanent bilateral recurrent laryngeal nerve palsy. The percentage of the patients who underwent recurrent laryngeal nerve excision decreased over time. The frequency of permanent recurrent laryngeal nerve palsy did not change significantly (Table 5), despite the increased use of subtotal or total thyroidectomy. Of 38 patients with permanent hypoparathyroidism, 17 (27%) were among the 63 patients who underwent total thyroidectomy, and 21 (2.0%) were among the 1035 patients who underwent subtotal thyroidectomy. There were no patients in whom hypocalcemia developed among the 1325 patients treated with thyroid lobectomy or partial lobectomy.
The 5-,10-, and 20-year recurrence-free survival rates were 94.9%, 90.0%, and 83.1%, respectively, for group 1 and 95.2%, 91.2%, and 83.9%, respectively, for group 2; for group 3 the rates were 96.2% at 5 years and 90.1% at 10 years. Compared with the 5-, 10-, and 20-year disease-specific survival rates for group 1 (98.4%, 95.5%, and 90.3%, respectively), the rates improved significantly (P=04) for group 2 (99.0%, 97.8%, and 93.9%, respectively); the rates also improved significantly (P=.03) for group 3 (5-year, 99.5%; 10-year, 98.2%) (Figure 3).
Because age at diagnosis and the size of the tumor are important prognostic factors in patients with papillary thyroid cancer, patients were divided into subgroups by age at diagnosis and tumor size as follows: (1) younger than 50 years, tumor size smaller than 25 mm (n=722); (2) younger than 50 years, tumor size 25 mm or larger (n=732); (3) 50 years or older, tumor size smaller than 25 mm (n=496); and (4) 50 years or older, tumor size 25 mm or larger (n=473). The 10- and 20-year disease-specific survival rates were more than 96% in the first 3 subgroups (no significant differences among the 3 groups). For patients in the last subgroup, the disease-specific survival rates were significantly better for groups 2 and 3 than for group 1 (Figure 4).
The sites of the first recurrence, of which approximately 80% were local recurrence in all 3 groups, are summarized in Table 6. Statistical analyses for the sites of recurrence were not done because of the different follow-up durations among the 3 groups. Follow-up durations varied among the 3 groups. To compare these groups, we analyzed using the information within 5 years after surgery. There were no significant differences in the first site of recurrence among the 3 groups.
The results of univariate and multivariate analysis of risk factors are shown in Table 7 and Table 8, respectively. In the univariate analysis, age, sex, tumor size, extracapsular invasion, and gross nodal metastases were identified as significant risk factors for group 1, whereas sex did not seem to influence the disease-specific survival for group 2.
In the multivariate analysis, age, sex, tumor size, and gross nodal metastasis remained significant predictors of survival for group 1; however, only age and gross nodal metastasis were significant for group 3. The extent of surgery was not included in the univariate and multivariate analysis because, as reported by Noguchi et al,7 survival was not affected by the type of thyroid surgery performed on 2192 patients with papillary carcinoma who were treated January 1, 1965, to December 31, 1988. In the multivariate analysis in all groups, the group is a significant independent risk factor (Table 8), indicating that the improvement of the disease-specific survival was caused by the improvement of the treatment.
In the present study, we excluded the patients with small thyroid cancer (tumor size less than 10 mm), because these tumors usually are found incidentally in surgical specimens, and ordinarily the patient has an excellent prognosis. Noguchi et al16 recently reported a long-term follow-up of 867 patients with small carcinomas of the thyroid, of whom 733 (84.5%) had papillary carcinomas; only 2 patients died of the disease after surgery during the mean follow-up period of 12.8 years.16
In almost all studies, the age of the patient at diagnosis and the size of the tumor are significant prognostic risk factors for patients with differentiated thyroid cancer.2- 7 In the present study, the most obvious changes in the characteristics of the groups were those in patient age and tumor size, both of which significantly influenced the prognosis. The mean age and the percentage of the patients aged 50 years or older at diagnosis were significantly increased; conversely, the mean diameter of the tumor significantly decreased (Table 1). The recent advances in diagnostic modalities and greater public awareness about thyroid cancer may have contributed to the earlier diagnosis experienced by groups treated more recently. The older age at diagnosis can be explained by the increased longevity in recent years rather than by delay in diagnosis.
Although no randomized studies comparing the risks and benefits among the different operative procedures have been reported, those who favor thyroid-conserving procedures cite the low incidence of recurrence and fewer complications of this type of treatment,7,11,13,17,18 while those advocating total thyroidectomy note the advantages in postoperative monitoring, radioactive iodine treatment, and the removal of undetected contralateral malignant neoplasms.5,19- 22 Our standard practice for papillary carcinoma has been subtotal thyroidectomy with modified radical neck dissection of the affected side. The extent of thyroidectomy and lymph node dissection, which are 2 main points of controversy among endocrine surgeons, were previously described in detail by analyzing a large number of patients treated at our clinic.7,23 The extent of thyroidectomy changed during the 3 periods; the percentage of subtotal and total thyroidectomy progressively increased, whereas that of partial lobectomy progressively decreased. Consequently, the incidence of recurrence at the remnant thyroid also tended to be reduced. Improvement of the preoperative diagnosis accounted for the decline in the percentage of performance partial lobectomy.
We observed an improvement of the disease-specific survival rates for patients in groups 2 and 3 compared with those in group 1, suggesting that the patients were treated at an earlier stage in groups 2 and 3 (Table 1 and Figure 1). The improvement of preoperative diagnosis of papillary thyroid cancer could contribute to the selection of a surgical procedure that is more appropriate for the patient. By analysis of patients' age and the size of the tumors, we noted a significant improvement in the disease-specific survival only among patients who were 50 years old or older with tumors 25 mm or larger in the maximum diameter (Figure 4). Thus, patients with anticipated poorer outcomes because of larger tumors experienced improvements in treatment, while for those with a good prognosis, no statistically significant difference was evident. A bias in lead time is possible because of the earlier diagnosis of thyroid cancer in recent years. However, this possibility was excluded by multivariate analysis of significant risk factors, including the group (Table 8). The failure to find significant improvement in the recurrence-free survival rate may merely reflect the earlier detection of recurrent diseases with the increasing use of sensitive ultrasonography, thallium scintigraphic scanning, and serum thyroglobulin measurement. This result also can be interpreted to mean that the patients were treated at an earlier stage of recurrent disease in groups 2 and 3 than were the patients in group 1.
We did not observe a significant improvement in the disease-specific or the recurrence-free survival rate for the patients in group 3 compared with patients in group 2 despite the introduction in our institution of sensitive imaging modalities at the beginning of the period represented by group 3. For patients in whom recurrences developed, the same percentages of patients were treated for a benign thyroid nodule in groups 2 and 3. This result suggested that improved diagnosis after the introduction of more sophisticated modalities did not directly affect the outcome. Furthermore, the diagnostic modalities were used for follow-up examination for patients in group 2: recurrent thyroid cancer developed in 122 patients in group 2, but the diagnosis of recurrence for 90 (73.7%) of the 122 patients was made during the period 1983-1990. Care should be taken when assessing the new diagnostic or therapeutic modalities, as they also influence the outcome of previously treated patients, especially patients with papillary thyroid cancer, because a considerable number of patients with recurrent thyroid cancer can be cured with surgery.
In the univariate and multivariate analyses, age was found to be the most important prognostic factor irrespective of the period represented by the group; this finding is in agreement with those cited by many authors.1,3- 7,10,11,13,22,24 In the multivariate analysis, age, sex, tumor size, and gross nodal metastasis were the independent significant risk factors in group 1, whereas only age and gross nodal metastasis were significant in group 3. The interpretation of these results is limited because of the different follow-up durations for the 3 groups. However, we can consider sex to be a nonsignificant risk factor at least for the patients in group 2 because the 14-year mean follow-up is long enough for the analysis. Some authors have stressed the importance of sex as one of the major determinants of survival in patients with papillary thyroid cancer,7,13,25 whereas others have reported converse findings.4,5 This discrepancy could be attributable to the difference in patient cohorts from one institution to another.
The presence of gross nodal metastases was identified as one of the important risk factors for all groups. Tubiana et al,25 Scheumann et al,24 and other authors6,13,22,26,27 have reported that the presence of clinical nodal metastases was associated with decreased survival rates and with increased recurrence rates as well. However, some studies have shown that the presence of nodal metastases has no adverse effect on survival3,4,13 or may be associated with improved prognosis.11 One explanation would be that patients with clinically palpable metastases underwent more extensive surgical procedures, resulting in a prognosis comparable to that for patients with no noticeable metastasis. Alternatively, the results depend on the percentage of younger patients included in the population studied, because younger patients more often have nodal metastases despite the better prognosis. In the univariate analysis in the present study, tumor size was not significant for group 3. For this result, more follow-up seems necessary before a definitive conclusion is drawn.
Postoperatively, the patients were given a thyroid hormone (levothyroxine sodium) for replacement rather than for prophylactic suppression during the follow-up period because of a lack of convincing data of a beneficial effect of thyrotropin suppressive therapy and its adverse effects on bone and cardiac function, especially in elderly and postmenoposal women. Although no randomized trials of thyrotropin suppressive therapy as an adjunct to surgery have been conducted, Mazzaferri and Jhiang5 and others28 showed a beneficial effect of such treatment. However, Cady et al29 reported that the administration of thyroid hormone for differentiated thyroid cancer to suppress thyrotropin did not improve survival time. Unless distant metastasis developed and iodine accumulation in the metastatic lesion was confirmed, postoperative radioactive sodium iodide I 131 (Radiocaps-131) ablation was not performed at our institution. Patients with distant metastases at presentation were excluded from the present study. The main reason for not using radioactive sodium iodide prophylactically is the limited capacity for the therapy because of rigid regulations for the use of radioisotopes in Japan. In addition, recurrence from distant sites is rare. The effects of radioiodine and thyrotropin suppressive therapy were not analyzed in the present study.
The present study showed that the risk factors for papillary cancer changed during a long period because of differences in patient characteristics and the diagnostic and therapeutic modalities used. Because similar changes may prevail at other institutions, the results derived from risk factor analysis should be considered in the treatment of and prediction of survival for patients with papillary thyroid cancer.
Reprints: Hiroyuki Yamashita, MD, Noguchi Thyroid Clinic and Hospital Foundation, 6-33 Noguchi-Nakamachi, Beppu Oita 874, Japan (e-mail: email@example.com).