Oncological Outcome Prediction in Differentiated Thyroid Carcinoma: Assumption or Improved Accuracy?

Accuracy in the prediction of adverse oncological outcomes has been a challenge in endocrine neoplasia, where a complex web of factors including host-dependent variables and genomic-epigenomic signatures tends to shape tumor behavior. ^{1 –3} Several scoring schemes and evidence-based guidelines have been generated to facilitate dynamic risk stratification and are often used to rationalize the need for adjuvant radioiodine therapy in patients with differentiated thyroid cancer (DTC). ^2,3 In this issue of Thyroid, Wijewardene et al. introduced a novel predictive model for DTC that derives from their assessment of clinicopathological variables. ⁴

The proposed predictive model modifies and extends the risk stratification model of the American Thyrois Association (ATA) based on a large cohort with the addition of extrathyroidal extension (ETE) and stimulated thyroglobulin (sTg) as independent predictors of both tumor recurrence and synchronous metastasis (defined as distant metastasis at time of the DTC diagnosis on structural imaging or identified at time of the initial post-treatment whole body iodide scan). ⁴

In their retrospective study, the authors created a four-tiered ETE status based on histological criteria. These include category 1 (intrathyroidal tumor), category 2 (tumor that underlies the junction of the thyroid parenchyma and adjacent soft tissue, called the “pseudocapsule,” and no extension into soft tissue planes), category 3 (microscopic growth into adjacent perithyroidal fibroadipose tissue), and category 4 (widespread extrathyroidal spread into soft tissue and/or extension into skeletal muscle including the strap muscles). ⁴

Historically, the definition of minimal ETE has included tumor invasion of fibroadipose tissue or adipose tissue surrounding the thyroid parenchyma. ⁵ However, it is well documented that fat and fibroadipose tissue can be present within the thyroid and that the thyroid “pseudocapsule” is incomplete and inconsistent, ⁵ making the interpretation of such invasion unreliable, and resulting in lack of significance of this finding. This ultimately led the American Joint Committee Cancer (AJCC) to change the criteria for ETE to require strap muscle invasion documented histologically and confirmed by gross inspection (intraoperative, radiological, or macroscopic examination) to define pT3b disease in the 8th edition of the AJCC staging scheme. ⁶ Indeed, accuracy in defining ETE requires careful histological and anatomic confirmation. ⁶

For instance, a reactive desmoplastic tissue reaction is often indistinguishable from genuine ETE using the naked eye. However, microscopic strap muscle invasion in the absence of gross invasion has been lost in the 8th edition of the AJCC and should be reintroduced as a risk factor in DTC. In the current series, the distinction between categories 2 and 3 may not be reproducible, since their illustrated criteria do not account for normal anatomic variations in many patients. Unlike the authors' assumption, normal thyroid parenchyma and related cellular proliferations often extend into perithyroidal soft tissue, and intrathyroidal fat can be trapped within intrathyroidal tumors, giving rise to inappropriate distinctions between categories 2 and 3.

Some exceptions would be the involvement of subcutaneous adipose tissue, which is usually seen in association with strap muscle involvement in lateral lobes or muscle involvement in the isthmus where Sommering's muscle is intrathyroidal in normal circumstances. These anatomical variations necessitate the collection of confirmatory data.

The criteria applied to diagnose vascular invasion (angioinvasion) have also been inconsistent, ⁷ which may explain lack of correlation with biological outcome in some series (as discussed in Mete and Asa ⁷ ). In the field of thyroid carcinoma, the distinction of angioinvasion from lymphatic invasion has been well recognized as essential, given the different prognosis of patients with locoregional lymph node spread due to lymphatic involvement when compared with distant metastases associated with angioinvasion. The 2022 WHO classification of thyroid tumors has endorsed this requirement as a standard of practice. ³

In the article by Wijewardene et al., the authors indicate that they adopted strict criteria to define angioinvasion, but they report difficulty to distinguish lymphatic invasion from capillary vessel invasion. In this series, the data analyses were based on the presumption that most papillary thyroid carcinomas with small vessel invasion had lymphatic invasion, and most follicular thyroid carcinomas or oncocytic carcinomas with small vessel invasion had capillary-type angioinvasion. This assumption should ideally have been confirmed using immunohistochemical biomarkers to distinguish lymphatic invasion from angioinvasion. ^7,8

Moreover, the criteria used to define angioinvasion even in large vessels are subject to interobserver variability. It seems intuitive that any study reporting a lack of association between recurrence/metastatic disease and angioinvasion should raise concern about the validity of the criteria applied.

The proposed predictive model endorses sTg as a predictive variable; while this is an enticing proposition, an increasing number of patients with early-stage DTC undergo lobectomy or hemithyroidectomy rather than total thyroidectomy, in which case the predictive role of sTg is unreliable.

Improved accuracy of the prediction of adverse oncological outcomes remains an important goal for those who diagnose and manage DTC. The four-tiered ETE criterion proposed by the authors is not without flaws, however, it does emphasize the importance of pathological variables. The addition of sTg is another controversial proposal, but it too confirms the importance of clinical and biochemical assessment in the management of this disease that is almost unique in the important role that sex and age play in prognosis.