Lobectomy and completion thyroidectomy rates increase after the 2015 American Thyroid Association Differentiated Thyroid Cancer Guidelines update

Introduction

Papillary thyroid carcinoma (PTC) is the most common subtype of differentiated thyroid cancer, which comprises over 90% of all thyroid cancers (Sherman 2003, Haugen 2016). PTC has the best prognosis of all thyroid cancer subtypes (Haugen 2016). We examine the impact of the American Thyroid Association’s ‘2015 Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer’ (2015 ATA Guidelines) (Haugen 2016) on the treatment of PTC in a tertiary-referral endocrine surgery unit (the Unit).

There has been a recent trend towards more conservative management of PTC, based on population studies which demonstrated no difference in disease-specific survival between thyroid lobectomy (TL) and total thyroidectomy (TT) for low-to-intermediate risk patients (Mendelsohn et al. 2010, Adam et al. 2014). This is reflected in the 2015 ATA Guidelines, which permit either TL or TT for PTCs of 1–4 cm without evidence of extra-thyroidal extension or nodal metastases (eligible PTCs) (Haugen 2016). In contrast, previous guidelines recommended TT for PTCs ≥1 cm whilst reserving TL for unifocal PTCs <1 cm (Cooper 2009).

Due to the indolent nature of PTC, there have been no prospective randomised trials assessing the surgical management of eligible PTCs (Papachristos et al. 2021). Instead, the 2015 ATA Guidelines are based on large retrospective analyses such as Bilimoria et al. (2007) and Adam et al. (2014) which utilised population-level data from the National Cancer Database. Since the publication of the 2015 ATA Guidelines, a re-analysis of the same data by Rajjoub et al. (2018) found that patients with PTC from 2 to 4 cm performed worse when treated with TL vs TT, which highlights the difficulty in interpreting retrospective population data. This uncertainty has recently been acknowledged in recent guidelines from the American Association of Endocrine Surgeons, recommending that initial management of PTCs of 1–4 cm should be individualised (Patel et al. 2020).

The 2015 ATA Guidelines suggest that TL alone is sufficient for ‘low-risk’ eligible PTCs that do not require radioiodine therapy (Haugen 2016). Proponents of this approach argue that TL minimises the risk of complications (Welch & Doherty 2018) and may obviate the need for life-long thyroxine replacement (Chaker et al. 2017). However, there is an increased risk of local recurrence and a need to monitor the contralateral lobe (Mazzaferri & Jhiang 1994, Haugen 2016). Furthermore, the utility of serum thyroglobulin as a sensitive tumour marker to detect disease recurrence is lost.

The primary difficulty in deciding which patients are suitable for TL is the inability to perform accurate pre-operative risk stratification. Pathological features, such as vascular invasion and lymph node metastases, often prompt reclassification to the ‘intermediate risk’ category under the ATA’s modified 2009 Risk Stratification System, which would ordinarily warrant adjuvant treatment with RAI and thus require completion thyroidectomy (CT) (Haugen 2016). Indeed, we have previously evaluated the rate of discordance between pre-operative criteria for lobectomy and definitive post-operative risk stratification and postulated it to be as high as 43.5% (Haugen 2016, DiMarco et al. 2019). The aim of this study is to examine the real-world rate of TL and CT with reference to the 2015 ATA Guidelines.

Materials and methods

A retrospective cohort study of consecutive patients undergoing surgical intervention for PTC at the Endocrine Surgery Unit of a single academic tertiary-referral centre was undertaken. We aimed to assess the impact of the 2015 ATA Guidelines on the rates of TL, CT, surgery for local recurrence and major complications in adult patients with eligible PTCs. Ethics approval to access the de-identified data was obtained from the Northern Sydney Local Health District Research Governance Office (2020/ETH02787).

All adult patients surgically treated for PTC through the Unit from January 2013 to March 2020 were considered for inclusion. De-identified data were collected from a prospectively maintained endocrine surgery database, including patient age, sex, pre-operative fine-needle aspirate (FNA) cytology, type of operation, surgical complications and post-operative histopathology. Immediate complication data are routinely recorded for all patients at the time of admission. Post-operative PTH is routinely measured on all patients undergoing CT or TT, and all patients with temporary hypoparathyroidism are followed up until it resolves, or until they are classified as having permanent hypoparathyroidism. Permanent hypoparathyroidism is defined as a need for ongoing calcium or calcitriol supplementation at 12 months post-operative and includes patients with partial recovery of PTH levels.

Patients considered eligible for TL as definitive surgical management under the ATA Guidelines were included for analysis. Specifically, these patients had (i) Bethesda V/VI FNA cytology; (ii) tumour diameter >1 and ≤4 cm and (iii) no pre-operative evidence of extrathyroidal extension or lymph node metastases (Haugen 2016). Post-operatively measured tumour size was used as a proxy for pre-operative tumour size. All patients underwent pre-operative ultrasonographic assessment of the central and lateral compartments, and inclusion was contingent on a clinical N0 nodal status.

Patients were divided into two cohorts, determined by whether their initial operation was before or after the publication of the 2015 ATA Guidelines: January 2013–December 2015 (pre-publication) and January 2016–March 2020 (post-publication).

For each cohort, we compare the rate of TL and TT as a percentage of primary procedures. We also describe the rates of CT among patients initially treated with TL, re-operative surgery for local recurrence and major complications (defined as permanent hypoparathyroidism, permanent recurrent laryngeal nerve palsy, re-operation for haemorrhage, wound infection or death). Where the initial surgical procedure occurred in the pre-publication period, any follow-up procedures (including CT and re-operation for local recurrence) were also counted in the pre-publication period. The initial treatment episode included index operation and any subsequent procedures planned in the immediate post-operative period. Local recurrence was defined as patients who required a follow-up lymph node dissection alone, after completion of their initial treatment episode, or patients with FNA-confirmed pathological lymph nodes managed conservatively during the follow-up period. Patients with contralateral lobe recurrence were also considered to have recurrent disease.

Patient age, sex, tumour size and FNA cytology were compared between the pre- and post-publication periods using an independent two-sample t-test in the Real Statistics Resource Pack for Microsoft Excel. All statistical tests were two-sided, with a significance level of 0.05. Primary and secondary outcomes were compared between the pre- and post-publication periods using a χ–² test (Yates 1934, Swinscow 1997). Fisher’s exact test was used when the smallest expected value was <5 (Cochran 1954).

Results

Patient demographics and tumour characteristics

A total of 461 cases met inclusion criteria for the study: 156 pre-publication and 305 post-publication (Fig. 1). As described in Table 1, there was no significant difference in age, sex, tumour size or pre-operative cytology between the pre-publication and post-publication periods.

View Full Size

Case selection.

Citation: Endocrine Oncology 3, 1; 10.1530/EO-22-0095

• Download Figure
• Download figure as PowerPoint slide

Table 1

Pre- and post-publication population characteristics.

	All eligible PTCs (n = 461)	Pre-publication (n = 156)	Post-publication (n = 305)	Independent two-sample t-test
Age in years, mean (s.d.)	49 (14)	49 (14)	48 (15)	t(459) = 0.27, P = 0.78
Sex, n (%)				t(459) = 0.60, P = 0.55
Female	368 (80%)	127 (81%)	241 (79%)
Male	93 (20%)	29 (19%)	64 (21%)
FNA cytology, n (%)				t(459) = 0.50, P = 0.62
Bethesda V	135 (29%)	48 (31%)	87 (29%)
Bethesda VI	326 (71%)	108 (69%)	218 (71%)
Tumour size in mm, mean (s.d.)	19 (7.4)	20 (7.6)	19 (7.3)	t(459) = 1.3, P = 0.20
Tumour size category, n (%)
10 mm < x ≤ 20 mm	307 (67%)	99 (64%)	208 (68%)
20 mm < x ≤ 30 mm	114 (25%)	40 (26%)	74 (24%)
30 mm < x ≤ 40 mm	40 (8.7%)	17 (11%)	23 (7.5%)

Completion thyroidectomy rate

In the pre-publication period, three of seven patients treated with TL required CT (43%). This compared to 21 of 56 (38%) patients managed initially with TL in the post-publication period (P = 1.0). The reasons for CT were consistent between the two periods, as depicted in Fig. 2. Post-operative features leading to CT were multifactorial in the majority (63%) of cases, involving a combination of nodal involvement, local invasion or lymphovascular invasion.

View Full Size

Reasons for completion thyroidectomy.

Citation: Endocrine Oncology 3, 1; 10.1530/EO-22-0095

• Download Figure
• Download figure as PowerPoint slide

Notably, of the 249 total thyroidectomies performed for 'eligible PTCs' in the post-publication period, 159 (64%) had intermediate or high-risk features on post-operative histopathology that would have necessitated CT, had TL initially been performed. The reasons for this post-operative risk re-categorisation in these cases are set out in Table 3. There were no cases where distant metastases led to post-operative risk re-categorisation.

Table 3

Reasons for post-operative risk re-categorisation of eligible PTCs managed primarily with total thyroidectomy.

Results of post-operative histopathology	Number	Percentage
Remains ‘low’ risk	90	36.1
Upstaged to ‘intermediate’ or ‘high’ risk	159	63.9
Lymph node metastases	129	51.8
Local invasion	91	36.6
Lymphovascular invasion	77	30.9
Total	251	100

Discussion

This study describes the real-world impact of the introduction of the 2015 ATA Guidelines on decision-making in the surgical management of low-risk PTC at an endocrine tertiary referral centre. After the introduction of the 2015 ATA Guidelines, the rate of TL for eligible PTCs increased significantly; however, the majority (82%) of eligible PTCs were still treated with TT. Our findings are similar to observations based on the Surveillance Epidemiology and End Results (Sutton et al. 2022) and National Surgery Quality Improvement Program databases (Ullmann et al. 2019). The modest engagement with the ATA recommendations can be partially attributed to the high incidence of multinodular goitre in the area in which the group practice, which contributes significantly to the fact that TT is frequently recommended as the initial treatment for patients with PTC highlights. However, the results also underscore the complexity of the decision-making process, through which a variety of disease factors, patient factors and clinician factors must be considered before committing to an operative approach.

Among the 56 (18%) patients treated with TL, 21 (38%) required CT based on post-operative histological findings. This finding reinforces the work of Hirshoren et al. (2018), DiMarco et al. (2019) and Kuo et al. (2021) in highlighting the inaccuracy of our currently available pre-operative risk stratification systems. The pathological indication for upstaging to intermediate or high risk, and recommendation of CT were multifactorial in the majority of patients. The subtleties of interpretation of the clinical and pathological data, informed by clinical experience in a multidisciplinary forum, mean that the application of the guidelines to real-world practice is inevitably shaped by clinician judgement, rather than strict adherence. Interpretation of risk associated with certain pathological findings may be associated with inter-observer variability. For example, the distinction between lymphatic and vascular invasion is often not possible to make because it is difficult to differentiate morphologically small lymphatics from small vascular spaces on haematoxylin and eosin slides. Ultimately, the decision to perform CT is driven by the post-operative risk assessment and need for RAI treatment and so the RAI treatment philosophy of the treating team is an important determinant of the viability of TL as an initial surgical approach (Papachristos et al. 2021). Hence, we acknowledge that institutional practices may vary with respect to the interpretation of the guidelines, and ongoing longitudinal analysis of outcomes and recurrence data is important in evaluating the impact of their introduction.

No pre-operative risk factor has been shown to predict the need for CT (DiMarco et al. 2019). Authors such as Kluijfhout et al. (2017), Dhir et al. (2018) and DiMarco et al. (2019) predicted an increase in TL would translate to an increase in completion thyroidectomies because high-risk features such as local invasion and lymphovascular invasion – which may be indications for TT to facilitate RAI treatment – cannot be assessed pre-operatively. Our real-world data, demonstrating that approximately 40% of eligible patients initially treated with TL will ultimately require CT, confirm previously published theoretical calculations. This contributes valuable information to the preoperative informed consent discussion and ensures patients understand the relative likelihoods of the potential management pathways. Many patients may accept the risk of a staged procedure given the 60% chance of avoiding TT.

We did not observe any significant change in local recurrence despite a modest increase in lobectomy rates. However, our study was not powered to assess differences in rates of local recurrence. Furthermore, PTC is an indolent tumour and rates of local recurrence are low even at 10 years (Grant & Hay 1988). The median follow-up time was significantly lower in the post-publication period (2.9 years) than in the pre-publication period (6.9 years), resulting in potential follow-up time bias.

Kandil et al. (2013) reported that TL is less risky than TT, even in the hands of high-volume surgeons (Hauch et al. 2014, Haugen 2016) and this contributes to the drive to de-escalate surgical management. Due to the excellent prognosis of PTC regardless of management algorithms, any complications that occur must be evaluated with the utmost scrutiny. In our series, the overall complication rate was less than 1% and similar regardless of the operative approach. We acknowledge, however, that in non-specialist units the rates of permanent recurrent laryngeal nerve injury and permanent hypoparathyroidism may be as high as 1.5% (Sosa et al. 1998) and 12.5% (Annebäck et al. 2020, Koimtzis et al. 2021), respectively. Therefore, given more than half of thyroidectomies are performed by surgeons who perform fewer than five thyroidectomies per year, these risks must be carefully weighed in the decision to perform TT vs TL.

This study is limited by its retrospective design and associated potential for statistical bias in the collection and interpretation of retrospective data. Secondly, patients were included according to post-operative tumour size despite ultrasound being used for preoperative decision-making. Ultrasound tends to overestimate size (particularly for microcarcinomas <1 cm), meaning that some PTCs may have been excluded from our analysis because they were <1 cm post-operatively, despite being treated as 1–4 cm based on pre-operative ultrasound (Hahn et al. 2016). We excluded 205 microcarcinomas with a median size of 8 mm that would have been eligible PTCs if >1 cm.

Our analysis was also confined to histopathologically confirmed PTC. However, pre-operative decision-making is based on FNA which gives a likelihood of malignancy (60–75% for Bethesda V and 97–99% for Bethesda VI) (Cibas et al. 2009). Therefore, some lesions which would have been treated as PTCs on the basis of FNA may have been excluded from our analysis because they turned out to be benign on post-operative histopathology.

Conclusions

While the TL rate in this cohort of 'low-risk' PTC patients remains low, we observed a significant increase after the publication of the 2015 ATA Guidelines. Approximately 40% of patients with eligible PTCs treated with TL ultimately require CT. Patients should be counselled regarding the likelihood of this outcome during the informed consent process.

Declaration of interest

The authors declare that they have no relevant or material financial interests that relate to the research described in this paper.

Funding

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Author contribution statement

The authors confirm contribution to the paper as follows: Study conception and design: MS, BW; Data collection: AA, AP; Analysis and interpretation of results: MS, AP, BW; Draft manuscript preparation: AP, BW, AA, AG, SS, RCB, DL, VT, MG, BR, MS. All authors reviewed the results and approved the final version of the manuscript.