Patient and treatment demographics
shows the characteristics of the 522 enrolled PTC patients. Though we generally tried to keep the indications for thyroid lobectomy in ATA [1
] and NCCN [2
] guidelines, there were some additional inclusions in our study. One hundred ninety-three patients (37.0%) over 45 years of age and with incidentally detected 11 (2.1%) multiple thyroid cancer were included. The median age was 41.0 years, and 17.2% (92 patients) were male. If gross extrathyroidal extension was suspected, immediate conversion to total thyroidectomy was performed and the patient was excluded from this study; there were 55 (10.5%) microscopic capsular invasion and 16 (3.1%) isthmic resection margin invasion. Median tumor size was 0.5 cm (range, 0.1 to 4.5 cm). The median follow-up duration was 42.0 months (range, 3.0 to 93.0 months). Ipsilateral central neck dissection following thyroid lobectomy was done on high risk patients, but not prophylactic in all patients. Central node dissection was performed in 258 (49.4%), and there were 69 (13.2%) N1a, 182 (34.9%) N0, and 283 (54.2%) Nx stage patients after lobectomy ().
Description of patients under thyroid lobectomy (n = 522)
Immediate completion thyroidectomy
Of the 522 patients excluding 10 follicular and two medullary carcinomas, 34 (6.5%) underwent completion thyroidectomy for the remaining thyroid, central, and lateral neck recurrence. There were 21 immediate (61.8%) and 13 delayed (38.2%) completion thyroidectomies (). Immediate completion thyroidectomy was defined as occurring within three months after initial lobectomy. Immediate completion thyroidectomy was performed due to missed PTC (6 cases), positivity of margin or lymphovascular invasion (5 cases), disease free but required additional radioactive iodine therapy (15 cases) and remaining central metastases (10 cases).
Follow-up of 522 patients underwent lobectomy due to papillary thyroid cancer. CLND, central lymph node dissection.
Time course after lobectomy due to papillary carcinoma
In the past, we recommended routine completion thyroidectomy immediately or within three months after initial thyroid lobectomy in high risk patients. However, some patients desired observation over a period of time, even though this offered the possibility of relapse. The desire for and proportion of observational study was higher in the minimal invasive type compared with the classic type or endoscopic approach lobectomy, although this was not statistically significant (P = 0.183).
and shows the patterns of central or lateral recurrence in the two arms for each of the three nodal groups after thyroid lobectomy, Follicular and medullary carcinoma patients were excluded from further analysis. In the observation arm of lobectomy Nx, there were four thyroid and two central recurrences, and in the observation arm of N0, there were five (2.8%) delayed completions. Three recurred papillary thyroid carcinomas on remaining thyroid developed in the 15th, 27th, and 57th months. Three (1.7%) central recurrences developed in the 9th, 23rd, and 57th months. In immediate completion arm of N1a, there was no recurrence. But in the observation arm of lobectomy N1a, there were two delayed completions (3.3%) and confirmed central recurrences in the 11st and 33rd months later after initial lobectomy. A total of six thyroid, five central, and two lateral recurrences were observed in a median of 42.0 months. Lateral recurrences occurred in the 46th month after immediate completion in the N0 group and in the 30th month after immediate completion in the Nx group, but did not occur in the N1a observation arm ().
Fifteen recurrence after thyroid lobectomy
Comparison of observation arm of postlobectomy N1a and N0 groups
In four arms of N1a and N0 patients, two observation arms of N1a and N0 patients were further analyzed. Age (P = 0.938), sex (P = 0.188), tumor size (0.63 cm vs. 0.60 cm, P = 0.647) and follow-up period were not significantly different. The N1a observation arm (n = 61) was statistically equivalent to the N0 observation arm (n = 180) for central or lateral recurrence (). We assumed central metastasis after thyroid lobectomy and higher ratio of positive nodes over harvested node to be factors for central or lateral recurrence, as recommended in the ATA [1
] and NCCN [2
] guidelines. But the ratio of the harvested central nodes did not show a significant difference between the immediate completion arm versus the observation arm within N1a groups (62.5% vs. 67.7%, P = 0.583).
Observation without completion thyroidectomy with N1a versus N0 groups in papillary thyroid cancers
displays the Kaplan-Meier survival curves for central or lateral recurrences between observation arms in the N1a versus N0 group; it has not been adjusted for tumor size less than 1.0 cm. Even by excluding subgroup analysis for tumors 1 cm or larger, no significant difference was found between the N1a versus N0 observation arms (P = 0.365) within the 42.0-month follow-up period.
Kaplan-Meier survival curves for central or lateral recurrences according to observation arm of N1a versus N0 groups (P = 0.365, log-rank test).
Risk evaluation for central or lateral recurrence
We used Cox proportional hazards modeling of the central or lateral recurrence after thyroid lobectomy (). Indeterminate contralateral thyroid tumor masses were identified on retrospective radiologic review, so remaining thyroid recurrences were excluded from the event of recurrence.
Cox proportional hazard analysis of variables predicting central or lateral neck recurrence after thyroid lobectomy (n = 391)
Using data from 399 PTC patients with full details on both central and lateral recurrences, age over 45 years old (HR, 1.54; 95% confidence interval [CI], 0.24 to 9.92; P = 0.652), male (HR, 0.92; 95% CI, 0.09 to 9.06; P = 0.943), and TSH suppression vs. levothyroxine replacement after one year TSH suppression (HR, 0.58; 95% CI, 0.87 to 3.80; P = 0.566) were not significant. Surprisingly, the N1a observation arm was not associated with a significantly worse prognosis for recurrences compared with the N0 observation arm (HR, 0.33; 95% CI, 0.04 to 2.77; P = 0.308 in univariate analysis / HR, 0.58; 95% CI, 0.06 to 5.69; P = 0.643 in multivariate analysis).