Comparative analysis of pre‐operative ultrasound and histopathology in small papillary thyroid carcinoma in the era of active surveillance

Contemporary management of small papillary thyroid carcinomas (PTC) includes active surveillance (AS) as a number of these tumours are indolent. Overseas studies have reported AS in tumours up to 15 mm. This study aims to look at an Australian cohort of patients who have had surgery for non‐incidental PTCs and analyse their pre‐operative ultrasound and histopathology data to investigate potential issues that might arise in the era of AS.


Introduction
The prevalence of small papillary thyroid cancers (PTCs) is increasing largely due to widespread use of various radiological modalities. [1][2][3] Most of these tumours are indolent, many are slow to progress or do not progress at all, and the long-term survival rate for low-risk PTCs is 99%. [3][4][5][6][7][8][9][10][11][12][13][14] Active surveillance (AS) has been advocated as a treatment option for small PTCs with the aim of avoiding unnecessary intervention, potential surgical risks and thyroxine replacement. [4][5][6][8][9][10][15][16][17] AS first appeared in the 2015 American Thyroid Association (ATA) Guidelines for Adult Patients with Thyroid nodules and Well Differentiated Thyroid Cancer and a framework for managing patients was reported by Brito et al. in 2016. 5,18,19 There are other factors that need to be considered in deciding whether a patient is suitable for AS. Many patients do not feel comfortable living with a diagnosis of carcinoma and some patients may not want to participate in an AS program that requires long term review with repeat investigations. The balance between surgery or AS requires consideration of all these factors. In addition, the preoperative framework proposed by Brito et al. does not and cannot take into consideration important pathological variables that are used to determine appropriate risk stratification.
This study aims to investigate the importance of having this pathological information by comparing the pre-operative ultrasound of PTCs ≤15 mm to the histopathological data and assess whether histopathological risk factors are present that may influence decisions regarding suitability for AS.

Methods
Retrospective study undertaken with the Sydney Head and Neck Cancer Institute database, Australia (ethics approval number X16-0367). Cohort initially comprised of patients who had undergone surgical intervention for PTC with a primary tumour size ≤15 mm between 2009 and 2019 (n = 306). Patients that had nonincidental carcinoma or a neck dissection were excluded. Two hundred and twenty-four patients were also excluded primarily based on unavailability of their pre-operative ultrasound scans (n = 215) or incidental carcinomas (n = 9). A final cohort of 82 patients were analysed.
The data points for this study were based on the framework guidelines from Brito et al. 5 The factors were split in four main domains: demographics, ultrasonography, histopathology and follow up (see Table 1). For radiology data, an experienced sonologist reviewed the pre-operative scans for each patient. Once data collection was complete, patients were deemed either ideal, appropriate or inappropriate for AS based on Brito et al. 5 The framework focuses on ultrasound, patient, and medical team characteristics. For our research, medical team characteristics were not studied. Low-risk PTCs were defined as having no lymphovascular invasion (LVI), no perineural invasion (PNI) and no or minimal extrathyroidal extension (ETE). In the analysis, data between ultrasonography and histopathology findings were compared to identify any correlations or discrepancies.

Results
The demographics of the cohort of 82 patients are shown in Table 2. Female patients comprised 81% of the cohort. The median age range for each gender was similar (50 years males and 45 years females).
Fifty patients (61%) underwent an initial total thyroidectomy, with 30 patients (36%) having a hemithyroidectomy and 2 patients an isthmusectomy. Nine of the patients who had initial hemithyroidectomy went on to have a completion thyroidectomy. Only one of these nine patients had a single focus of a 0.5 mm carcinoma in the completion thyroidectomy specimen. Five of the nine patients who had a completion thyroidectomy were deemed appropriate for AS based on their pre-operative ultrasound.
On pre-operative ultrasound, the range of size of tumours was 5-15 mm (median 10 mm). The difference between the ultrasound measurement of tumour size and histopathology tumour size was within 2 mm in 57 patients (69.5%) of patients. The range of discrepancy of the tumour size between ultrasound and histopathology varied from 1 to 7.5 mm. With respect to ETE, ultrasound reporting of ETE correlated with histopathological ETE in 60 patients (73.2%) with a sensitivity of 18.5% and a specificity of 100%.
Based on the radiological framework set out in Brito et al., 68 patients (82.9%) would have been ideal or appropriate for AS, demonstrating that a large portion of patients with pre-operative ultrasound tumours ≤15 mm may not have required surgical intervention (Table 3).  Unifocal tumours were identified in 53 patients compared to multifocal disease in 29 patients (Table 2). Overall, tumour diameter ranged from 4 to 15 mm (median 10 mm). In our cohort, multifocal bilateral disease was only seen in tumours <10 mm. Unilateral multifocal tumours ranged from 0.2 to 10 mm (median 2 mm) and bilateral multifocal tumours ranged from 0.2 -9 mm (median 0.6 mm).
There were six identified variants of papillary thyroid carcinomas. The most prevalent type was classical followed by mixed follicular and papillary variant as shown in Table 2. There were no aggressive variants of PTC identified in our study.
Lymphovascular invasion was present in 29 patients (35.4%) ( Table 4). Twenty-five of these patients would have been eligible for AS based on their ultrasound characteristics. There was one patient who had PNI and was deemed appropriate for AS based on ultrasonography findings. Extrathyroidal extension status was divided into no ETE, minimal-ETE (into perithyroidal soft tissue) and gross-ETE (extension into strap muscles or trachea). 20,21 On histology, no ETE was reported in 55 patients (67.1%), minimal-ETE in 18 patients (22%) and gross-ETE in 9 patients (11%). Four of the nine patients with gross-ETE would have been deemed appropriate for AS based on their ultrasonography features.
BRAF V600E immunohistochemistry status was performed in 33/82 patients (25 positive and 8 negative). Fourteen of the BRAF status positive patients had a primary PTC diameter ≥10 mm.
Metastatic disease in perithyroidal lymph nodes was found in 17 patients (20.7%). The number of positive lymph nodes ranged from 1 to 4 and the size of metastatic deposits ranged from 0.2 to 8 mm (median 1 mm).
Based on the Brito et al. framework using pre-operative ultrasound findings, 83% of the patients in this study would meet criteria for AS. However, following review of the histopathology data, 59.8% of the patients had at least one pathological adverse factor.

Discussion
Small papillary thyroid carcinomas are becoming increasingly prevalent. Given many are often indolent and the survival rate of low-risk PTCs is 99%, less invasive management is deemed appropriate. In the 2015 ATA Guidelines for Adult Patients with Thyroid nodules and Well Differentiated Thyroid Cancer, AS as an option for low-risk PTCs was introduced. 19 In 2016, Brito et al. described a framework based on pre-operative factors that can help in the decision making of patients that might be selected for AS based on ultrasound findings, patient factors and medical team factors. 5 However, it is important to compare how the AS framework works with a cohort of patients that have undergone surgery to identify potential pitfalls or areas of concern. Our study has attempted to address this by comparing pre-operative ultrasound findings and histopathological findings in a cohort of patients undergoing surgery for PTCs ≤15 mm. The reason 15 mm (as opposed to 10 mm) was decided as the cut off for the inclusion criteria is based on more recent published literature from Japan on AS in tumours up to 15 mm in size. 14 In our study, the incidence of PTCs in females is higher compared to males which is consistent with published Australian data. 3 Total thyroidectomy was the initial surgery performed in 60.9% of the cohort which is not surprising as a large portion of the cohort had surgery prior to 2015 in an era where total thyroidectomy was routinely performed. Our results found that over half (58%) of these patients had unifocal disease and potentially could have been treated with hemithyroidectomy alone with reduced risk of hypoparathyroidism, potential injury to both recurrent laryngeal nerves and reducing the need for thyroxine replacement. 15,22,23 A metaanalysis published in 2021 demonstrated that total thyroidectomy was the main surgical management approach for low-risk PTCs between 2014 and 2019 which is the same time period as our data. 24 A review by McLeod et al. published in the Lancet journal in 2013, described how total thyroidectomy was favoured as the primary treatment option for low-risk PTC due to factors such as multifocal disease that is often occurs in both thyroid lobes. 22 The risk of multifocal disease (unilateral or bilateral) is sometimes used as an indication to proceed with surgical intervention, however multifocal papillary thyroid microcarcinoma without ETE is considered low-risk in the ATA Risk Stratification System.
The range and median diameter of the tumours was similar when measured on pre-operative ultrasound and histopathology. In most patients (69.5%) the discrepancy in the tumour size on ultrasound and histopathology was ≤2 mm. However, the higher end of discrepancy was up to 7.5 mm. Reasons for this discrepancy could be attributed to sonographer experience, placement of the ultrasound markers for measurement, nodule characteristics such as a variable cystic and solid component and shrinkage of tumours placed in formalin. Hahn et al. demonstrated that ultrasound accurately recorded the size of low-risk PTCs with a variation of 1 mm and that there was a 64.1% correlation of ultrasound size and pathology measurement, which is similar to the results seen in our study. 25 In our study, all the pre-operative ultrasound scans were reviewed by one sonologist with extensive experience in thyroid ultrasound. This may have induced a bias given the scans were reviewed by only one person and the person was highly specialized in reporting thyroid ultrasound. Despite this, the high degree of concordance between ultrasound and histopathology measurements gives reassurance that high resolution ultrasound closes matches histology size which is important when tumour size is used in the assessment of AS and in the clinical AJCC TNM staging system for thyroid cancer. None of the cohort had suspicious lymph nodes in preoperative imaging. What is apparent in patients undergoing AS is that important pathologically variables are not able to be determined unless surgery is performed. The rates are demonstrated in our study. These include aggressive pathological variants, lymphovascular invasion, perineural invasion, lymphovascular invasion and metastatic disease in central compartment lymph nodes. This information also allows for appropriate risk stratification that guides further treatment and prognosis. In our study, there were no aggressive variants, about 1/3 of patients had ETE (minimal 22% and gross 11%) and about 1/3 of patients had LVI. PNI was seen in one patient. 20% of the cohort had metastatic disease in perithyroidal lymph nodes (range 0.2-8 mm). Given that none of the cohort had abnormal lymph nodes on pre-operative ultrasound and central compartment lymph nodes with metastatic deposits >2 mm are ATA intermediate-risk stratification, this important pathological feature may be missed in patients undergoing AS based on ultrasound finding alone.
The presence of the BRAF V600E mutation is associated with a more aggressive PTC with higher rates of LVI and locoregional recurrence. [26][27][28] The role of BRAF V600E mutation in papillary thyroid microcarcinoma is unclear. 29 In our institution, BRAF V600E staining has only been routinely performed over the last 8 years which accounts for the large number of patients with missing BRAF V600E mutation status. As BRAF V600E positive status is used in the ATA Risk Stratification System, the number of patients deemed low-risk or intermediate-risk in our study might change. BRAF status is not discussed in the Brito et al. framework for AS paper and its role in AS is currently unknown.
Our study identified that 83% of our cohort were either ideal or appropriate for AS based on pre-operative ultrasound finding. This demonstrates that many patients potentially could have avoided surgical management if based on ultrasound criteria alone. However, following review of the histopathology data, 59.8% of the patients had at least one histopathological risk factor. This discrepancy is important and an area that needs further investigation to fully understand the implications of AS.
As we proceed in an era of AS of management of small well differentiated thyroid cancers it is important to consider that guidelines for AS cannot take into consideration histopathological risk factors that are used in clinical practice in determining decisions for further intervention such as completion thyroidectomy, radioactive iodine and TSH suppression. Despite this, AS is an established treatment option for many patients with small thyroid cancers. High resolution ultrasound with the addition of the TIRADS thyroid nodule reporting has assisted clinicians in making appropriate treatment recommendations. The challenge we face is ensuring that AS programs do not undertreat patients with thyroid carcinoma. In addition, limitations of AS include the cost of prolonged follow up with repeat ultrasound imaging and the anxiety that some patients have with a diagnosis of cancer and the recommendation of no surgical intervention. Given the high curability of papillary thyroid carcinoma, the relevance of adverse pathological features and its impact is likely to be in disease free survival outcomes only.
The final sample size of this study was unfortunately significantly less than the number of patients that were identified from the SHNCI database initially. This was due to missing data and was mainly due to the inability to obtain pre-operative scans. This is one weakness of our retrospective study and would be improved in a prospective study design. However, it is still possible to draw conclusions from the 82 patients in this study.

Conclusion
This study has shown that a significant proportion of patients who would have been eligible for active surveillance based on Brito et al. criteria had at least one adverse pathology risk factor. Also, there are limitations to the use of ultrasound to stratify patients accurately when deciding whether active surveillance is feasible. Although active surveillance is a valid management option, its application needs to take into consideration the lack of pathology information that best defines low-risk patients. With time and with further prospective studies, the indications and outcomes of active surveillance programs will be better defined.