Abstract

The incidence of papillary thyroid carcinoma, which accounts for 80-90% of all thyroid cancers, has recently been increasing. The current study aimed to compare the oncological and functional outcomes of total thyroidectomy (TT) and thyroid lobectomy (TL). To this end, a retrospective single-centre cohort study involving a tertiary care institution was conducted. Data regarding demographics, clinicopathology and postoperative complications from 586 patients with papillary thyroid cancer treated in a single institution were collected. Cox proportional-hazards models were utilised to determine differences in outcomes strati- ed according to propensity score. Our data suggested no signicant difference in the risk for locoregional recurrence or distant metastasis between TL and TT among patients with pT1-2 pN0 papillary carcinoma. TT plays an important role in improving prognosis among patients with metastatic lymph nodes in the central neck compartment (pN1a) (p=0.001). Moreover, TT had signicantly higher rates of postoperative hypocalcaemia and recurrent laryngeal nerve paralysis compared to TL (p<0.001 and p=0.02, respectively).

Introduction

The incidence of papillary thyroid carcinoma (PTC), which accounts for 80-90% of all thyroid cancers, has currently been increasing due to increased diagnostic scrutiny and better diagnostic technology 1. The widespread availability and improved sensitivity of neck ultrasonography (US) and fine-needle aspiration cytology (FNAC), accounting partly as overdiagnosis, and effects of environmental and lifestyle changes, may be responsible for such an increase 2,3. The Surveillance, Epidemiology and End Results (SEER) database reported that incidence rates for PTC increased 2.4-fold during the few last decades predominantly among those with low-risk intrathyroidal T1-T2 carcinoma 4,5. Despite its high prevalence, thyroid cancer is rarely deadly with an excellent prognosis and a 10-year survival rate exceeding 90% 6-9. Unfortunately, disease recurrence and postoperative surgical complications remain very common 10. No universal agreement on the management of PTCs, especially the smallest lesions, has been established given that confidently differentiating between aggressive and slowly progressing diseases is not yet possible. To date, thyroid surgery has been the main treatment for PTCs. The extent of curative surgery has long been a controversial topic, with evidence for both conservative surgery (lobectomy) and radical surgery (total thyroidectomy). Although the 2009 American Thyroid Association (ATA) guidelines had recommended total thyroidectomy (TT) for PTCs > 1 cm, the recent 2016 ATA guidelines recommend lobectomy alone for low-risk, 1-4 cm PTCs 11,12. Recent studies have observed that surgery type did not affect patient outcomes and were much more related to cancer intrinsic risk factors or patients themselves 13-20. Other studies have reported that thyroid lobectomy (TL) had a higher risk for recurrence compared to TT 9,21,22. Furthermore, adverse surgical events, such as hypocalcaemia and recurrent laryngeal nerve injury, are more frequent and severe after TT than after lobectomy alone 23,24. As confirmed by the latest ATA guidelines, accurate preoperative staging and risk evaluation is crucial to determine the proper surgical approach and post-surgical management 12. While some potential risk factors, such as histological features, vascular invasion and extrathyroidal extension (ETE), are difficult to evaluate preoperatively, others, like age, sex, familial history, BRAFV600E mutation and lymph node (LN) involvement can be readily assessed preoperatively 25-30. To date, the role of surgical extension among patients with PTC remains controversial. As such, the present study aimed to compare long-term outcomes of TT and TL among patients with T1-T2 PTC and evaluate postoperative complication rates. To minimise selection bias, various statistical analyses were performed to categorise our cohort according to the presence or absence of lymphadenectomy (central neck dissection) while accounting for prognostic and confounding factors.

Materials and methods

Study population

This retrospective single-centre cohort study was approved by the European Institute of Oncology Ethics Committee. From January 1995 to January 2018, a total of 3,013 patients underwent surgical treatment at the Otolaryngology and Head and Neck Surgery Department of the European Institute of Oncology IRCCS, Milan, Italy. Patients (n = 2,424) with the following characteristics were excluded: age less than 18 years, benign disease, previous thyroid treatment, mixed-type PTC, tumour size > 4 cm, pT3-pT4 stage, cN1b, distant metastasis, follow-up < 6 months and contralateral TL completion performed for reasons other than suspected or proved recurrence. Ultimately, 586 patients with pT1 or pT2 PTC were enrolled and assigned to two groups according to whether they received TT (group; n = 412) or TL (group; n = 174). Patient demographics and surgical and histopathologic details (histological features, lesion size, multifocality and LN metastases) were determined from the IEO electronic archive. To minimise selection bias, we divided our cohort in two groups according to the presence of central neck dissection during surgical planning, considering that most locoregional recurrences (LRRs) involve locoregional cervical lymph nodes.

Patient management

Preoperative patient evaluation included a complete clinical examination, routine blood tests, laryngeal fibroscopy to determine vocal cord mobility, US to estimate size, possible ETE and LN metastases in the central area (VI level) and FNAC 31. Most cases underwent FNAC for preoperative diagnosis of PTC, whereas others were diagnosed incidentally through surgical specimens from patients undergoing treatment for a cytologically benign disease.

Surgical approach

TL was performed among patients with no evidence of ETE, no bilateral carcinoma and a lesion size of < 2 cm in diameter when FNAC was positive for carcinoma and > 2 and < 4 cm when preoperative FNAC was Tyr 2. TT was performed among cases with bilateral carcinoma, oncological lesion ≤ 4 cm, benign lesion > 4 cm, bilateral goiter, familial history of thyroid cancer and uncertain contralateral nodules. All surgical approaches were determined through multidisciplinary team decisions with the patient’s consent. Clinically node-negative patients underwent prophylactic central neck dissection when thyroid FNAC was Tyr 3, 4, or 5, as recommended by the multidisciplinary team, or when macroscopic ETEs were detected during preoperative US or following surgery 12,32-34.

The aforementioned surgical management was standardised at our Institute starting in 2010, during which a more efficient, minimally invasive surgery and systematic use of surgical loupes for all surgeons were introduced.

Radicalisation was proposed when central neck metastasis or ETE were discovered to allow for radioactive iodine treatment (RAIT) in TL cases.

Surgical complications and follow-up

An analysis of the two most frequent postoperative thyroidectomy complications (i.e., transient or permanent hypocalcaemia and transient or permanent recurrent laryngeal nerve paralysis) was performed. Only euthyroid patients with normal preoperative calcium and PTH levels were included. As suggested by American Association of Clinical Endocrinologists guidelines, hypocalcaemia was defined as a serum albumin-corrected total calcium level lower than 2.1 mmol/L (8.5 mg/dL) regardless of signs and symptoms 35. Permanent hypocalcaemia was defined as requirement of vitamin D and calcium supplementation for more than 12 months after surgery, independent of calcium values 35. Patients with aggressive cancer features and risk factors, such as pT2 stage and/or central node involvement, underwent RAIT according to current guidelines 12,33,34. The postoperative follow-up included clinical evaluation consisting of physical examination, laryngeal fibroscopy, neck US, iodine-131 scan (for TT) and serum thyroglobulin samples at 6- to 12-month intervals (for TT and TL). Outcomes evaluated included LRR, metastasis occurrences, and mortality. LRR was defined as a new lesion in the thyroidectomy bed or cervical LNs detected through clinical examination with neck US and increased thyroglobulin levels and confirmed using FNAC or histopathological samples from the re-intervention. Metastasis was detected using iodine-131 scans (for TT), computed tomography, or cytology and histology. Recurrence-free survival (RFS) was defined as the time interval, expressed in months, between the first surgery and detection of recurrence (both LRR and distant metastasis).

Statistical analyses

All analyses comparing the type of thyroidectomy were stratified according to the presence or absence of lymphadenectomy. Categorical variables were presented as relative frequencies (percentages). Continuous variables were reported as median and interquartile range. Categorical variables were compared using Fisher’s exact test, while continuous variables were compared using the Wilcoxon signed-rank test. The Kaplan-Meier method was utilised to generate survival curves, which were compared using the log-rank test. To reduce the impact of confounding factors and treatment selection bias, a propensity score was used to identify factors significantly associated with thyroidectomy type. Cox proportional-hazards models were stratified according to the propensity score. All tests were two-sided with a p-value of < 0.05 considered statistically significant.

Results

Patient characteristics and surgical technique(s)

Demographics and clinicopathological features of the 586 patients are presented in Table I. Patients had a median age of 48 years [interquartile range (IQR): 39-58 years], among whom 461 were women (78.7%). Multifocality was found in 182 patients (31.1%). Moreover, 494 patients had pT1 PTC (84.3%), while 92 had pT2 PTC (15.7%). Among the 586 patients who satisfied the selection criteria, 412 (70.3%) underwent TT and 174 (29.7%) underwent TL. The median follow-up duration was 58 months (IQR: 24-102 months). None of the patients developed distant metastasis. Significant differences in age [50 (41-59) vs 43 (36-52) years; p < 0.001], multifocality (37.6% vs 15.5%; p < 0.001) and central neck dissection (62.1% vs 47.7%; p = 0.0012) were observed between the TT and TL groups, respectively, as expected from the pre-surgical studies (Tab. I). No significant differences in sex and pT stage were observed between groups (p = 0.52 and 0.86, respectively). Central lymphadenectomy was performed in 339 patients (57.8%). Demographics and clinicopathological features of those who did and did not undergo central lymphadenectomy are summarised in Tables II and III.

Postoperative follow-up of patients who underwent central neck dissection

No deaths related to thyroid disease occurred among patients who underwent central neck dissection. A total of four patients (1.17%) developed local relapse (two with LN metastases and two in the contralateral lobe), among whom three underwent TL and one TT. Among those who underwent central lymphadenectomy, significant differences in multifocality (43.4% vs 13.3%; p < 0.001), central LN metastasis (42.2% vs 15.7%; p < 0.001) and age (p = 0.05) were found between the TT and TL groups (Tab. II). To evaluate survival curves of TT and TL groups, multivariate logistic models were created to determine propensity scores using prognostic factors that were significantly associated with type of surgery: multifocality, LN involvements and age (Tab. IV). Our analyses found that RFS rates differed significantly between the TT and TL groups [p = 0.041; hazard ratio (HR): 11.127, 95% confidence interval (CI): 1.117-110.8)] (Tab. V, Fig. 1). In particular, this difference was found only among patients with central LN metastasis (pN1a) and not among those with pN0 (p = 0.001 and 0.16, respectively) (Figs. 2A, B).

Postoperative follow-up of patients without central neck dissection

No deaths related to thyroid disease occurred among patients who did not undergo central neck dissection. A total of six patients (2.43%) developed local relapse (three with lateral neck metastases, two in the contralateral thyroid lobe and one with both lateral neck and contralateral thyroid lobe metastases), among whom three underwent TL and three TT. Among those who did not undergo central lymphadenectomy, significant differences in age (p ≤ 0.001) and year of surgery were observed between the TT and TL groups (p = 0.003) (Tab. III). To evaluate survival curves of TT and TL groups, multivariate logistic models were established to determine propensity scores using prognostic factors that were significantly associated with type of surgery: multifocality, year of surgery and age (Tab. IV). Our analyses found no significant difference in RFS between TT and TL groups (p = 0.73; HR: 1.364, 95% CI: 0.228-8.159) (Tab. V; Fig. 2C).

Differences in complications rates between groups

Serum calcium levels were collected in 419 patients after surgery. Using the previously established definition for hypocalcaemia, 243 (58%) and 22 patients (5.2%) were determined to have postoperative transient and permanent hypocalcaemia, respectively. Transient hypocalcaemia was significantly more common in the TT group than in the TL group (64.9% vs 19.1%; p < 0.001), whereas no significant difference in permanent hypocalcaemia was observed, perhaps due to the number of subjects evaluated (Tab. VI). More cases of transient recurrent laryngeal nerve paralysis were observed among those who underwent TT than those who underwent TL (p = 0.02; Tab. VI). Only two patients developed permanent recurrent laryngeal nerve paralysis, both of whom belonged to the TT group (Tab. VI). One patient needed tracheostomy, which was removed at 4 months after surgery for movement in one vocal cord. The second one reported unilateral permanent vocal cord paralysis without respiratory problems.

No upper laryngeal nerve damage was noted, while none of the patients reported swallowing disorders due to impaired sensitivity during bolus passage.

Discussion

Despite the increasing incidence of PTC, mortality rates remain constant, while prognosis remains excellent 4.

The excellent behaviour of this tumour has led to ongoing debates regarding the necessity of surgical management 5,8. Two large studies by Bilimoria et al. (2007) and Adam et al. (2014) had reached opposite conclusions. Accordingly, Bilimoria et al. found that among 52,173 patients, those with PTC > 1 cm who underwent conservative surgery exhibited increased rates of recurrence 9. However, that study had several limitations mainly due to the use of administrative registries from which clinicopathological variables were missing. Conversely, Adam et al., who examined 61,775 patients with 1-4 cm PTC, found that OS was similar between patients who underwent conservative and radical surgery after multivariable adjustment for clinical and pathological factors 14. Supporters of TT underline the lower risk of recurrence, possibility of thyroglobulin administration during follow-up, and use of I-131 for both diagnostic and therapeutic aims 36. On the other hand, lobectomy has been preferred for its lower postsurgical morbidity and the absence of lifelong replacement therapy 23,24,37. Moreover, supporters of lobectomy have stated that several studies found no difference in recurrence rates and that patients who develop recurrence can still be successfully treated without increasing disease-specific mortality 15. The latest ATA guidelines on thyroid carcinoma management concluded that lobectomy alone may be sufficient for low-risk PTC of 1-4 cm12. This indicates that thyroid cancer management must be tailored to each patient according to both stage and risk factors 18-20,25-30. This approach challenges the predominant role of tumour size in the management of PTC 38,39. Moreover, studies have proposed that an active surveillance approach toward managing microPTC (PTC < 1 cm) among those with very low-risk (old age, unifocal and well-defined thyroid nodule margins, and no signs of extrathyroidal extension and neck lymph node involvement on US) seems to be a safe and effective alternative to surgical resection 40,41.

The results of our study substantiate the hypothesis that TL does not promote greater risk of recurrence/LRR or distant metastasis among patients with pT1-2 pN0 papillary carcinoma. Our findings showed that TT had a significantly higher rates of postoperative hypocalcaemia and recurrent laryngeal nerve paralysis than TL.

Moreover, 22 patients determined to have permanent hypocalcaemia need continuous replacement therapy to maintain serum calcium levels in the lower reference range. Unfortunately, we have no data on the incidence of osteoporosis and osteoporotic fractures in permanent hypocalcaemia. However, it is not known if the incidence could be influenced by other factors, such as BMI, heredity, sex, age, other conditions, or medical therapies. This should be considered together with an estimate of the financial costs of both treatments given the increasing interest in cost estimation nowadays. A cost-effectiveness analysis conducted at our institution revealed that TL and TT had a direct cost of € 3,167 (2,800-3,200) and € 5,099 (4,880-5,200), respectively 42. Several limitations of the current study need to be noted. First is the retrospective nature of the analyses with potential confounding variables and treatment selection bias. We note that prospective studies on this disease are very difficult to conduct given the extended life expectancy, extensive follow-up duration, costs associated with such a study and related ethical issues. Another limitation is the small sample size, as well as the short median follow-up duration. On the other hand, this is a single-centre cohort study with complete and accurate medical records. Moreover, the current study utilised propensity scores to reduce the bias of confounding variables related to treatment effect evaluation arising from simply comparing outcomes among those who did and did not receive treatment. This method attempts to do in a retrospective study what randomisation does in a prospective study. TT, on the other hand, plays an important role in the management of metastatic lymph nodes in the central compartment (pN1a), which leads us to highlight the importance of preoperative study to avoid staging errors and reduce the number of subsequent TLs for oncological radicalisation.

Conclusions

The present study found that TL did not increase the risk for LRR or distant metastasis among those within pT1–2 pN0 PTC and had no effect on RFS. Moreover, TT promoted significantly higher rates of postoperative complications, such as postoperative hypocalcaemia and recurrent laryngeal nerve damage, compared to TL. Additional studies with larger sample sizes and longer follow-up duration would be useful to confirm our data. Although a randomised prospective study would provide much greater statistical power, it may not be feasible considering the characteristics of the disease.

Figures and tables

Figure 1..

Figure 2..

Table I..

Lobectomy (LT) Total Thyroidectomy (TT) Total P-value
174(100) 412(100) 586(100)
pT 1 146 (83.9) 348 (84.5) 494 (84.3) 0.8653
2 28 (16.1) 64 (15.5) 92 (15.7)
Sex Females 134(77) 327 (79.4) 461 (78.7) 0.5244
Males 40(23) 85 (20.6) 125 (21.3)
Multifocality No 147 (84.5) 257 (62.4) 404 (68.9) < 0.0001
Yes 27 (15.5) 155 (37.6) 182 (31.1)
Year of surgery ≤ 2010 122 (70.1) 249 (60.4) 371 (63.3) 0.0263
> 2010 52 (29.9) 163 (39.6) 215 (36.7)
Age < 48 117 (67.2) 189 (45.9) 306 (52.2) < 0.0001
≥ 48 57 (32.8) 223 (54.1) 280 (47.8)
Lymphadenectomy No 91 (52.3) 156 (37.9) 247 (42.2) 0.0012
Yes 83 (47.7) 256 (62.1) 339 (57.8)
Prognostic factors according to type of surgery.

Table II..

Lobectomy (LT) Total Thyroidectomy (TT) Total P-value
83(100) 256(100) 339(100)
pT 1 77 (92.8) 214 (83.6) 291 (85.8) 0.037
2 6 (7.2) 42 (16.4) 48 (14.2)
Sex Females 59 (71.1) 194 (75.8) 253 (74.6) 0.393
Males 24 (28.9) 62 (24.2) 86 (25.4)
Multifocality No 72 (86.7) 145 (56.6) 217(64) < 0.0001
Yes 11 (13.3) 111 (43.4) 122(36)
Year of surgery ≤ 2010 42 (50.6) 137 (53.5) 179 (52.8) 0.644
> 2010 41 (49.4) 119 (46.5) 160 (47.2)
Age * < 48 56 (67.5) 142 (55.5) 198 (58.4) 0.054
≥ 48 27 (32.5) 114 (44.5) 141 (41.6)
pN1a No 70 (84.3) 148 (57.8) 218 (64.3) < 0.0001
Yes 13 (15.7) 108 (42.2) 121 (35.7)
Central lymphadenectomy group - prognostic factors according to type of surgery.

Table III..

Lobectomy (LT) Total Thyroidectomy (TT) Total P-value
91(100) 156(100) 247(100)
pT 1 69 (75.8) 134 (85.9) 203 (82.2) 0.046
2 22 (24.2) 22 (14.1) 44 (17.8)
Sex Females 75 (82.4) 133 (85.3) 208 (84.2) 0.555
Males 16 (17.6) 23 (14.7) 39 (15.8)
Multifocality No 75 (82.4) 112 (71.8) 187 (75.7) 0.056
Yes 16 (17.6) 44 (28.2) 60 (24.3)
Year of surgery ≤ 2010 80 (87.9) 112 (71.8) 192 (77.7) 0.003
> 2010 11 (12.1) 44 (28.2) 55 (22.3)
Age* < 48 61(67) 47 (30.1) 108 (43.7) < 0.0001
≥ 48 30(33) 109 (69.9) 139 (56.3)
No central lymphadenectomy group - prognostic factors according to type of surgery.

Table IV..

OR Low 95% CI High 95% CI P-values
Central lymphadenectomy Multifocality No vs yes 4.78 2.37 9.61 < 0.0001
pN No vs yes 4.30 2.18 8.49 < 0.0001
Age 0.97 0.95 0.99 0.0127
No central lymphadenectomy Multifocality No vs yes 2.30 1.14 4.65 0.0204
Age 0.93 0.91 0.96 < 0.0001
Year of surgery > 2010 vs ≤ 2010 0.34 0.16 0.74 0.0063
Multivariate logistic models to determine propensity scores used in the Cox models - prognostic factors significantly associated with type of surgery: association between hemi-thyroidectomy vs. thyroidectomy.

Table V..

HR Low 95%CI Up 95%CI P-values
Central lymphadenectomy 11.127 1.117 110.8 0.041
No central lymphadenectomy 1.364 0.228 8.159 0.733
Multivariate Cox proportional hazard models stratified by propensity scores - analyses are carried out in the two groups by lymphadenectomy.

Table VI..

Lobectomy (LT) Total Thyroidectomy (TT) Total P-value
174(100) 412(100) 586(100)
Parathyroidectomy No 165 (94.8) 357 (86.7) 522 (89.1) 0.0037
Yes 9 (5.2) 55 (13.3) 64 (10.9)
Vocal cord transient paralysis No 168 (96.6) 376 (91.3) 544 (92.8) 0.0233
Yes 6 (3.4) 36 (8.7) 42 (7.2)
Vocal cord permanent paralysis No 6 (3.4) 34 (8.3) 40 (6.8) 0.5541
Yes 0 (0) 2 (0.5) 2 (0.3)
Lobectomy (LT) Total Thyroidectomy (TT) Total p-value
63(100) 356(100) 419(100)
Transient hypocalcaemia No 51 (80.9) 125 (35.1) 176(42) < 0.0001
Yes 12 (19.1) 231 (64.9) 243(58)
Persistent hypocalcaemia Yes 0 (0) 22 (6.2) 22 (5.2) 0.2434
Complications by type of surgery.