2023 Volume 70 Issue 9 Pages 901-908
Active surveillance (AS) is an accepted management option for patients with low-risk papillary thyroid microcarcinoma (PTMC), although some patients undergo immediate surgery (IS). At surgery, patients may have risky features such as adhesion or invasion to the adjacent organs. The surgical outcomes of this subset of patients are unknown. Here, we investigated the surgical and oncological outcomes of these patients in comparison with others. Between 2005 and 2019, 4,635 patients were diagnosed with low-risk PTMC at our institute. Of these, 1,739 patients underwent IS. In total, 114 patients had risky features at surgery (risky feature group), while the remaining 1,625 did not (no-risky feature group). The median follow-up periods in the risky and no-risky feature groups were 8.5 and 7.6 years, respectively. The risky feature group had higher incidences of tracheal invasion (8.8%), recurrent laryngeal nerve invasion (RLN) (7.9%), and permanent vocal cord paralysis postoperatively (10.0%) and higher frequency of pathological lateral lymph node metastasis (6.1%) than the no-risky feature group (0%, 0%, and 0.2%, and 0%, respectively [p < 0.01]). However, unexpectedly, the former had a lower incidence of high Ki-67 labeling index (1.1%) and lower locoregional recurrence rate (0%) than the latter (8.3% and 0.7, respectively [p < 0.01], not calculable). None of the groups developed distant metastasis or died of the disease. The risky feature group required resection of the trachea and/or the RLN more often than the no-risky feature group. However, unexpectedly, the tumor growth activity in the risky feature group was low, and their oncological outcome was excellent.
ACTIVE SURVEILLANCE (AS) for cT1aN0M0 papillary thyroid carcinoma (PTC) and specifically low-risk papillary thyroid microcarcinoma (PTMC) was initiated at Kuma Hospital (Kobe, Japan) in 1993 and at the Cancer Institute Hospital (Tokyo, Japan) in 1995 [1]. Owing to its favorable outcomes, this management strategy was accepted as an alternative to surgical treatment and is starting to be implemented worldwide [2-8]. However, this strategy should not be applied in patients with PTMC who have definite high-risk clinical features such as apparent lymph node metastasis, distant metastasis (although rare), evidence of aggressive cytology (rare), vocal cord paralysis due to carcinoma invasion to the recurrent laryngeal nerve (RLN), or evident invasion to the trachea on imaging studies [9].
However, PTMC with no such high-risk clinical features are not always indicated for AS. Indeed, tumors not attached to the trachea or not located on the dorsal side of the thyroid are ideal for AS. However, careful evaluation using ultrasonography and computed tomography (CT) is required for tumors attached to the trachea or located on the dorsal side of the thyroid. In 2016, we revisited 1,143 PTMCs with no high-risk features which underwent surgery and showed that 24% of tumors measured ≥7 mm, attached to the trachea at obtuse angles between the tumor surface and tracheal cartilage, showed tracheal invasion requiring tracheal resection, and 9% of these having no normal thyroid tissue between the tumors and course of the RLN invaded the nerve requiring partial layer resection or complete dissection (with reconstruction) of the nerve [10]. Patients with these features on imaging studies are regarded as risky for AS and generally recommended surgical treatment [11, 12]. Although patients showing invasion to the trachea or RLN at surgery may be expected to have poor surgical and oncological outcomes, to the best of our knowledge, there are no reports on the outcomes of this subset of patients. Therefore, in this study, we investigated the clinicopathologic characteristics and the surgical and oncological outcomes of the patients with low-risk PTMC who were reported with risky features at surgery.
The indications of immediate surgery for PTMC without clinical high-risk features are 1) patients’ preference and 2) having the possibility of invasion to the trachea or RLN. Regarding the latter, at present, we evaluated the invasion to these organs according to the criteria described in the introduction; obtuse angles formed by the tracheal cartilage and tumor surface and lack of normal thyroid tissue between the surface of the tumor located on the dorsal side of the thyroid and the course of the RLN are regarded as risk factors for invasion to the trachea and RLN, respectively. However, in the past, the evaluation was based on each physician’s estimation and their method of judgment on the possibility of carcinoma invasion to adjacent organs for most cases was unknown. Therefore, in this study, we defined PTMC with no high-risk features as mentioned in the introduction as a low-risk PTMC.
Enrollment of patientsFor this study, we enrolled the same patient cohort with low-risk PTMC as that in our previous studies [13, 14]. Between 2005 and 2019, 4,635 patients were diagnosed with low-risk PTMC (cT1aN0M0) based on cytology (Bethesda V or VI) (Fig. 1). None of them had high-grade cytology variants such as the tall cell variant. Tumor status and locoregional nodal status were evaluated with ultrasound examinations, and other imaging studies were added if indicated. Suspected nodes were evaluated with both fine needle aspiration cytology and measurement of thyroglobulin in the washout of the needle. We diagnosed lymph nodes as metastasized when they were cytologically diagnosed as carcinoma metastasis or their thyroglobulin levels in the washout were high [15]. Of these patients, 2,896 opted for active surveillance and 1,739 (37.5%) underwent surgical treatment within 1 year after the diagnosis. During the surgery, 106 patients had risky surgical features such as adhesions or cancer invasion to the trachea and RLN. However, due to the lack of clear preoperative evidence of organ invasion, these tumors were classified as cT1a. Eight patients had small lymph nodes suspected of PTC metastasis on ultrasound examinations; however, the results of the fine needle aspiration studies were negative and the washout thyroglobulin levels were low. Thus, these 114 patients were classified as cT1aN0M0 and into the risky feature group. The remaining 1,625 patients did not have these risky features at surgery and were classified into the no-risky feature group.
Flow diagram presenting recruitment and classification of the patients with low-risk papillary thyroid carcinoma in the present study. Risky features at surgery and inappropriate for active surveillance include tumor adhesion or invasion to the recurrent laryngeal nerve, the trachea and other nearby tissues, and suspected, but not cytologically confirmed, regional lymph node metastasis. PTMC; papillary thyroid microcarcinoma.
Pre- and post-operative evaluations included blood tests for thyroid stimulating hormone (TSH), free T4, thyroglobulin, thyroglobulin antibody, calcium, and intact parathyroid hormone (PTH) (if indicated), and laryngoscopy. In the present series, all patients underwent laryngoscopy before surgery and were confirmed to have functioning bilateral vocal cords preoperatively. A postoperative laryngoscopy was also performed in all patients. If vocal cord paralysis was detected, we periodically repeated laryngoscopy and regarded patients as having permanent vocal cord paralysis when paralysis persisted at least one year after surgery.
Histopathological studiesSurgical specimens were post-operatively evaluated as a standard pathological procedure, and the cell proliferation activity of the tumor was evaluated by immunohistochemistry of the Ki-67 labeling index (LI) as described previously [16]. In this study, we performed Ki-67 immunohistochemistry for 1,247 cases in no-risky feature group and 89 cases in risky features group. Ki-67 immunohistochemistry was not performed for other cases mainly because of strong calcification. We observed hot areas under 400× magnification and calculated the percentage of Ki-67 positive cells. The threshold was set at 5% and divided into two categories: high (>5%) and low (≤5%). This was evaluated by pathologists at our institution.
Post-operative follow-upPatients were advised to regularly visit our clinic at least once a year. We performed blood thyroid tests and imaging studies (mainly ultrasound) during checkups. We performed cytological examinations for newly appeared suspected nodules in the remnant thyroid and suspected lymph nodes in the lateral neck. For the evaluation of lymph nodes, we also measured thyroglobulin levels in the washout of the needles used for cytology. We classified them as carcinoma recurrence when they were diagnosed as PTC or metastatic PTC. We sent yearly questionnaires to patients who were referred to other hospitals to follow up on their conditions. The follow-up periods (median and range in years) after surgery in the no-risky feature and risky feature groups were 7.6 (0.6–16.0) and 8.5 (1.0–14.9), respectively. There was no significant difference in the follow-up periods after surgery between the two groups (p = 0.63).
Study designThis was a retrospective, single-center, cohort study.
Statistical analysesStatFlex (Artec, Osaka, Japan) was used for chi-squared test and Mann-Whitney U test for comparing variables. For time-series survival analyses, Kaplan-Meier method and log-rank test were adopted. A p-value <0.05 was regarded as statistically significant.
Statement of human rightsThis study was approved by the Ethics Committee of Kuma Hospital (No. 20200709-1). The requirement of informed consent from the patients was waived due to the retrospective nature of the study.
First, we compared the baseline characteristics and preoperative findings of patients in the no-risky feature and risky feature groups. As shown in Table 1, the age at surgery and tumor size were higher in patients in the risky feature group than those in the no-risky feature group. The incidences of coexisting chronic thyroiditis and Graves’ disease were higher in the no-risky feature patients.
| Variables | No-risky feature group (N = 1,625) |
Risky feature group (N = 114) |
p-value | |
|---|---|---|---|---|
| Sex | Male | 182 (11.2%) | 11 (9.6%) | p = 0.61 |
| Female | 1,443 (88.8%) | 103 (90.4%) | ||
| Age | Median (ranges) | 55 (20–92) | 60 (24–78) | p < 0.01 |
| Chronic thyroiditisa | Yes | 530 (32.6%) | 23 (20.2%) | p < 0.01 |
| Graves’ disease | Yes | 129 (7.9%) | 3 (2.6%) | p < 0.05 |
| Family history of PTCb | Yes | 60 (3.7%) | 1 (0.9%) | p = 0.18 |
| Multiplicity at diagnosis | Yes | 180 (11.1%) | 18 (15.8%) | p = 0.13 |
| Tumor sizec at the PTC diagnosis (mm) | Median (ranges) | 8.0 (2.5–11.0) | 9.0 (3.5–10.0) | p < 0.01 |
Abbreviations: PTC, papillary thyroid carcinoma
Data are expressed as frequency (percentage) or as median (range) as indicated.
a Positive for anti-thyroglobulin antibody and/or thyroid peroxidase antibody
b One or more first-degree relatives had PTC.
c Evaluated with ultrasound examination
Table 2 shows the intraoperative findings of the two groups. The extent of thyroidectomy did not differ between the two groups. However, eight patients (7.0%) in the risky feature group also underwent lateral node dissection since lateral node metastasis was suspected on ultrasound (but not cytologically diagnosed and thyroglobulin levels in the wash-out were low). In the risky feature group, 58 patients (50.9%) showed tumor extension (adhesion and/or invasion) to the trachea, and 10 of these (8.8%) required window resection with airtight tracheocutaneostomy or one-stage suture of the wedge resection; five were pathologically diagnosed as invasion to the ligament between the tracheal cartilages (but not to the endotracheal membrane). Further, carcinoma extension to the RLN was detected in 51 patients (44.7%) and nine (7.9%) required resection and reconstruction of the nerve using the ansa cervicalis; four invaded to the RLN and the remaining five invaded to the surrounding tissue, but not to the axon on pathology [15]. None of the patients in the no-risky feature group showed carcinoma extension to adjacent organs.
| Variables | No-risky feature group (N = 1,625) |
Risky feature group (N = 114) |
p-value | |
|---|---|---|---|---|
| Extent of thyroidectomy | Total thyroidectomy | 716 (44.1%) | 52 (45.6%) | p = 0.75 |
| Lobectomy | 909 (55.9%) | 62 (54.4%) | ||
| Lymph node dissection | Central only | 1,625 (100.0%) | 106 (93.0%) | p < 0.01 |
| Central and Lateral | 0 (0.0%) | 8 (7.0%) | ||
| Suspected of lateral node metastasisa | Yes | 0 (0.0%) | 8 (7.0%) | p < 0.01 |
| Organs where carcinoma extendsb | Trachea | 0 (0.0) | 44 (41.1) | p < 0.01 |
| RLN | 0 (0.0) | 37 (34.6) | ||
| Trachea and RLN | 0 (0.0) | 14 (13.1) | ||
| Others | 0 (0.0) | 12 (11.2) | ||
| Extension to the trachea | Yes | 0 (0.0%) | 58 (50.9%) | p < 0.01 |
| Tracheal resection | No | 1,625 (100.0%) | 56 (49.1%) | p < 0.01 |
| Shaving only | 0 (0.0%) | 48 (42.1%) | ||
| Resection of all layersc | 0 (0.0%) | 10 (8.8%) | ||
| Extension to the RLN | Yes | 0 (0.0%) | 51 (44.7%) | p < 0.01 |
| RN resection | No | 1,621 (99.8%) | 63 (55.3%) | p < 0.01 |
| Shaving only | 0 (0.0%) | 42 (36.8%) | ||
| Accidental injuryd | 4 (0.2%) | 0 (0.0%) | ||
| Resection and reconstruction | 0 (0.0%) | 9 (7.9%) | ||
Abbreviations: Central, central lymph node dissection; Lateral, lateral lymph node dissection; RLN, recurrent laryngeal nerve.
Data are expressed as frequency (percentage).
a Suspected on ultrasound but not diagnosed on cytology.
b Based on intraoperative findings.
c Two patients underwent airtight tracheocutaneostomy and eight underwent suture repair after resection.
d Two cases underwent end-to-end anastomosis after accidental amputation.
Table 3 presents postoperative unfavorable events in patients in the two groups. Transient and permanent vocal cord paralysis occurred more frequently in patients from the risky feature group than in patients from the no-risky feature group. Eleven patients (10.0%) in the risky feature group showed permanent vocal cord paralysis due to carcinoma extension, while four (0.2%) in the no-risky feature group showed this complication as a result of accidental injury of the RLN. The incidence of transient/permanent hypoparathyroidism, hematoma requiring reoperation, and hypothyroidism did not differ between the two groups.
| Variables | No-risky feature group (N = 1,625) |
Risky feature group (N = 114) |
p-value | |
|---|---|---|---|---|
| Transient vocal cord paralysis | Yes | 141 (8.7%) | 30 (26.3%) | p < 0.01 |
| Permanent vocal cord paralysisa | Yes | 4 (0.2%) | 11 (10.0%) | p < 0.01 |
| Accidental injury | 4 (0.2%) | 0 (0.0) | p = 0.60 | |
| Due to tumor invasion | 0 (0.0%) | 11 (10.0%) | p < 0.01 | |
| Transient hypoparathyroidism | Yes | 383 (23.6%) | 24 (21.1%) | p = 0.54 |
| Permanent hypoparathyroidismb | Yes | 43 (2.6%) | 2 (1.8%) | p = 0.77 |
| Postoperative hematoma requiring re-operation | Yes | 12 (0.7%) | 1 (0.9%) | p = 0.59 |
| Levothyroxine administration after surgeryc | Yes | 331 (20.4%) | 24 (21.1%) | p = 0.86 |
Data are expressed as frequency (percentage).
a Persistent vocal code paralysis revealed by laryngoscopy one year after surgery or longer.
b Administration of vitamin D and/or calcium preparation one year after surgery or longer.
c Patients who were administered levothyroxine before surgery were not included.
As shown in Table 4, of the eight patients who underwent lateral node dissection due to the presence of suspicious nodes on imaging studies (although fine needle aspiration results were negative and thyroglobulin levels in the wash-out were low), seven were pathologically positive for lateral node metastasis. Ki-67 LI could be evaluated for 1,247 and 89 patients in the no-risky feature and risky feature groups, respectively. Only one tumor (1.1%) in the risky feature group showed high Ki-67 LI, and the incidence was significantly lower than that in the no-risky feature group, at 8.3% (p < 0.01). Recurrence in contralateral lobe and regional lymph nodes was detected in 0.4% and 0.2% of the no-risky feature group patients, respectively. In contrast, none of the patients in the risky feature group showed locoregional recurrence (p not calculated because the latter had 0 incidence). Fig. 2 shows the Kaplan-Meier curve of locoregional recurrence rate of the two groups; 10-year recurrence rate of patients in the no-risky feature and risky feature groups was 0.7% and 0.0%, respectively. No significant difference was observed between them. To date, none of the patients in either group showed distant recurrence or died of thyroid carcinoma.
| Variables | No-risky feature group (N = 1,625) |
Risky feature group (N = 114) |
p-value | |
|---|---|---|---|---|
| Pathological node metastasis | pN0 | 1,179 (72.6%) | 75 (65.8%) | p < 0.01 |
| pN1a | 446 (27.4%) | 32 (28.1%) | ||
| pN1b | 0 (0.0%) | 7 (6.1%) | ||
| Ki-67 LI | High (>5%) | 104 (8.3%) | 1 (1.1%) | p < 0.01 |
| Low (≤5%) | 1,143 (91.7%) | 88 (98.9%) | ||
| Unknown | 378 | 25 | ||
| Locoregional recurrence | Contralateral lobe | 6 (0.4%) | 0 (0.0%) | |
| Lymph node | 3 (0.2%) | 0 (0.0%) | ||
| Recurrence rate at 10 years | 0.7% | 0.0% | ||
| Distant recurrence | Yes | 0 (0.0%) | 0 (0.0%) | |
| Carcinoma death | Yes | 0 (0.0%) | 0 (0.0%) | |
Abbreviations: LI, labeling index
Data are expressed as frequency (percentage).
Locoregional recurrence rates of patients in the risky feature group and the no-risky feature group
This study compared backgrounds, clinicopathological features, adverse events, and oncological outcomes between low-risk PTMC (cT1aN0M0) patients with risky feature at surgery and those with no-risky feature at surgery.
In our series, 114 patients were regarded as the risky feature group due to carcinoma invasion to the adjacent organs, such as the trachea and RLN, being observed at surgery or lateral node metastasis being suspected on preoperative imaging studies (but not diagnosed as cytology and low levels of thyroglobulin in the wash-out). Most of these patients had preoperative imaging features suggestive, but not conclusive, of the involvement of adjacent organs. However, since clear evidence of invasion or metastasis was lacking based on imaging studies or cytology, they were still classified as cT1aN0M0. In the risky feature group, 10 patients (8.8%) required resection of all layers of the trachea and 9 (7.9%) required RLN resection with reconstruction using the ansa cervicalis [17]. Further, 8 patients (7.0%) were suspected of N1b on imaging studies, which indicated lateral node dissection. Accordingly, the extension of surgery was often broader in the risky feature group patients than in the no-risky feature group patients, causing a higher incidence of adverse events of surgery. In our series, although no significant difference in the incidence of hypoparathyroidism was detected between the two groups, transient and permanent vocal cord paralysis occurred more frequently in the risky feature group patients. Unfortunately, four (0.2%) of the 1,625 patients of the no-risky feature group suffered accidental injuries of the RLN, confirming the larger number of patients found by our previous report [9].
Patients in the risky feature group were significantly older, had larger tumors, and showed extrathyroid extension, which are predictors of poor prognosis on clinical PTC [18-20]. However, interestingly, in the present series of PTMC with risky surgical features, only one patient (1.1%) had a high Ki-67 LI and this incidence was significantly lower than that of the no-risky feature group (8.3%) (p < 0.01) (although it was still lower than clinical PTC, at 20% [21]). Previous studies showed that high Ki-67 LI significantly predicted poor oncological outcomes in clinical or larger papillary thyroid carcinoma patients [21, 22]. Moreover, in the PTMC subset, Hirokawa et al. demonstrated that in surgical specimens, Ki-67 LI was significantly higher in patients with PTMCs that enlarged during AS than in patients with non-enlarged PTMCs [16]. In the present study, Ki-67 LI studies indicated that patients with PTMCs in the risky feature group had lower growth activity than those in the no-risky feature group. This was an unexpected finding since the risky feature group was older, having larger tumors extending to the adjacent organs. However, the 10-year locoregional recurrence rate of the risky feature group was 0%, which was lower than that of the no-risky feature group (0.7%). We were unable to calculate the difference as the incidence of the risky feature group was 0.
Miyauchi et al. analyzed the natural history of low-risk PTMC and showed that hypothetical tumor growth activity before the presentation (assuming that a single carcinoma cell measuring 10 μm was present at birth) was significantly higher than that after presentation [23]. Ito et al. demonstrated that the growth activity of most PTMCs demonstrated a significant decrease after tumor enlargement [24]. These indicate that most PTMCs show rapid growth in early phase but the growth activity decreases thereafter. Further, as shown in Table 1, age in the no-risky feature group patients was significantly lesser than that in the risky feature group patients, and tumor size in the no-risky feature group was significantly lesser than that in the risky feature group. Considered together, we speculate that, although the tumors in the risky feature group once had a high growth activity and extended to adjacent organs, their growth activity might have subsequently decreased, which might explain the reason for higher Ki-67 LI in the no-risky feature group than that in the risky feature group.
The present data might raise a question on the indication for surgery in such patients. However, to date, no molecular markers have been identified to predict the growth of PTMC. Although the association of BRAF mutation and TERT mutation in clinical PTC patients was reported as a poor prognosis [25, 26], previous studies showed that TERT mutation is a rare event in patients with PTMC and is not useful as a prognostic marker [27-29]. Therefore, presently, surgical treatment is generally considered for most patients with PTMC belonging to the risky feature group.
Our study has some limitations. Our analysis was retrospective, and because most patients were first seen many years ago, the present grouping of no-risky feature and risky feature was based on surgical findings and not performed preoperatively based on the recent criteria [10, 11]. Furthermore, the median follow-up periods in our series were comparably short at 7.6 years for no-risk feature group and 8.5 years for risk feature group, respectively. Despite these limitations, our study showed that the growth activity of patients with PTMC in the risky feature group was generally low and their tumor characteristics was indolent.
In summary, PTMCs in the risky feature group often adhered to or invaded the adjacent organs such as the trachea and RLN or metastasized to the lateral nodes, requiring the resection of adjacent organs and resulting in a high incidence of adverse events. However, their growth activity was generally low and their oncological outcome was excellent. Although studies with long-term follow-up are necessary, local control might be more important than extensive surgery such as broad prophylactic node dissection and adjuvant therapy using high dose of radioactive iodine.
We would like to thank Mr. Makoto Kawakami for his contribution in extracting clinical data from the electronic medical record system and Editage (www.Editage.com) for the English language editing.
Author disclosure statementThe authors state that there are no conflicts of interest related to this study.
Funding statementNo funding was received for this study.
