Prevalence of Differentiated High-Grade Thyroid Carcinoma Among Well-Differentiated Tumors: A Systematic Review and Meta-Analysis

Abstract

Background:

The current edition of the World Health Organization (WHO) classification of endocrine tumors introduced grading for follicular cell-derived thyroid cancer. Tumors with necrosis and/or high mitotic count but not fulfilling the Turin criteria for poorly differentiated carcinoma will be reclassified as differentiated high-grade thyroid carcinoma (DHGTC). However, the impact of this reclassification has not been evaluated. In this study, we performed a systematic review and meta-analysis to estimate the prevalence of this new entry across thyroid tumor subtypes.

Methods:

In this systematic review and meta-analysis, studies reporting data on necrosis and/or mitoses in well-differentiated thyroid carcinoma (WDTC) were used to estimate the prevalence of DHGTC. Heterogeneity and potential publication bias were also evaluated. Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were followed, and quality assessment was performed using a modification of the Newcastle–Ottawa scale. The study has been registered in the International Prospective Register of Systematic Reviews (PROSPERO, ID: CRD42022378716).

Results:

In clinically unselected patients, the prevalence of DHGTC in WDTC was 0.072 [95% confidence interval, CI, = 0.045–0.113]. The proportion of high-grade tumors greatly varied across growth patterns and subtypes. Overall, the prevalence of DHGTC was higher in follicular thyroid carcinoma (FTC; 0.146 [CI = 0.101–0.205]) than in papillary thyroid carcinoma (PTC; 0.059 [CI = 0.036–0.097]). Diffuse sclerosing, follicular, and classic subtype PTC had the lowest rates of high-grade features (i.e., 0.018 [CI = 0.004–0.084]; 0.036 [CI = 0.010–0.124]; and 0.042 [CI = 0.027–0.066], respectively), while a greater proportion of solid trabecular and histologically aggressive PTC could be reclassified as DHGTC (i.e., 0.154 [CI = 0.067–0.314] and 0.168 [CI = 0.108–0.252], respectively). Similar proportions were obtained for minimally and widely invasive FTC (i.e., 0.136 [CI = 0.058–0.287] and 0.152 [CI = 0.086–0.254], respectively). Finally, in a cohort of patients with poor prognosis (i.e., fatal cases, metastatic and radioiodine resistant tumors, cases with biochemical recurrence), the proportion of DHGTC was 0.287 [CI = 0.155–0.469].

Conclusions:

Following the current WHO indications, some tumors will be reclassified as DHGTC. The proportion of tumors with high-grade features is relevant in FTC, solid trabecular, and histologically aggressive PTC subtypes. A remarkable enrichment in DHGTC among patients with poor prognosis confirms the negative impact of high-grade features on outcome.

Introduction

The fifth edition of the World Health Organization (WHO) classification of endocrine tumors was released at the end of 2022. One of the major changes was the recognition of two groups of high-grade non-anaplastic follicular cell-derived carcinomas: poorly differentiated thyroid carcinoma (PDTC) and differentiated high-grade thyroid carcinoma (DHGTC). PDTC was already present in the previous editions, and the criteria for its diagnosis date back to 2007.¹ The Turin criteria establish that PDTC has a poorly differentiated growth pattern (i.e., solid, trabecular, insular, or a combination of them), absence of nuclear features of papillary carcinoma, and at least one of the following features: convoluted nuclei, mitotic count ≥3 per 10 high-power field (HPF; now defined as 2 mm²), and tumor necrosis.

DHGTC represents a new entry in the WHO classification scheme, introduced due to the prognostic relevance of grading for tumors that—while retaining architectural and cytological characteristics of well-differentiated tumors—have unfavorable prognosis.² DHGTC retains the architectural and cytological characteristics of well-differentiated tumors but present high-grade features defined by the presence of tumor necrosis or mitotic count ≥5 per 2 mm².^3,4

The combined prevalence of PDTC and DHGTC ranges from 1% to 6.7% of all thyroid carcinoma.^3,5 Reports on necrosis and mitotic count in well-differentiated tumors have often been presented in aggregate with PDTC. Hence, the prevalence of high-grade features in well-differentiated carcinomas—and, consequently, the prevalence of DHGTC alone—is unclear. Similarly, the independent prognostic impact of necrosis and high mitotic count has been proven, but results were usually obtained in aggregate with PDTC defined according to the Turin consensus criteria.⁶ Recently, Wong et al. observed a poorer disease-free and disease-specific survival in papillary thyroid carcinoma (PTC) with high-grade features compared with PDTC.⁷

In this study, we performed a systematic review and meta-analysis to assess the prevalence of tumor necrosis and high mitotic count (i.e., ≥5 per 2 mm²/10 HPF) in well-differentiated thyroid carcinomas (WDTCs) to estimate the prevalence of this new entry in the current WHO classification scheme. We also performed a subgroup analysis by histological subtype.

Materials and Methods

Search strategy and selection criteria

A systematic search was performed on PubMed and Scopus databases following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.⁸ The strings used are provided in the Supplementary Data S1. The electronic search was performed on November 29, 2022, and updated on November 3, 2023. The following filters were applied: studies on humans, English-written articles, and publication date between January 1, 2003, and the day of search. Records identified from databases were merged and screened for duplicates using the automation tool available on Zotero 6 for Windows (www.zotero.org). Manual screening was then carried out to assess eligibility. Patients with WDTC were included, while those with anaplastic or poorly differentiated tumors as defined by the Turin consensus were excluded. Studies reporting tumor necrosis and/or mitoses assessed on histology were further considered.

In particular, data on mitoses were used only when the cutoff was ≥5 mitoses per 2 mm² (or per 10 HPF), or when the number of cases with mitotic count above the cutoff to determine differentiated high-grade thyroid tumor (i.e., ≥5 mitoses per 2 mm²) could be extracted. Studies reporting assessment of proliferative index by immunohistochemical markers (e.g., Ki-67, MIB-1) were not considered—since the use of immunohistochemistry markers is not explicitly recognized by the WHO scheme for diagnosing DHGTC of follicular cell origin—unless data on necrosis were provided. The complete Participants, Interventions, Comparisons, Outcomes, and Study design (PICOS) framework is reported in Table 1. The study was registered in the International Prospective Register of Systematic Reviews (PROSPERO, ID: CRD42022378716). The list of studies reviewed in full text and the reasons for exclusion are provided in Supplementary Table S1.

Table 1.

Eligibility Criteria Using Participants, Interventions, Comparisons, Outcomes, and Study Design Framework

Parameter	Inclusion criteria	Exclusion criteria
Population	Patients with follicular-derived well-differentiated thyroid tumors.	Patients with anaplastic or poorly differentiated tumors as defined by the Turin consensus.
Interventions	Tumors with necrosis and/or ≥5 mitoses per 2 mm² (or 10 high-power field) assessed on histology.	Mitoses assessed by IHC (e.g., Ki-67, MIB-1). Mitotic cutoff lower than defined in the inclusion criteria or no definition of cutoff.
Comparisons	Follicular-derived well-differentiated tumors without high-grade features (i.e., tumor necrosis and high mitotic count as defined in Interventions). Stratification by histologic subtype.
Outcomes	Prevalence of high-grade features and their impact on patient's outcome (i.e., disease-free survival, recurrence-free survival, overall survival, disease-specific survival).	Outcome not evaluated in the light of high-grade features. Outcome reported on aggregate with anaplastic or poorly differentiated tumors.
Study design	Original studies: cohort, cross-sectional, randomized trials, pre–post.	Case reports, narrative reviews, systematic reviews, and meta-analyses.

IHC, immunohistochemistry.

Data extraction

Record screening and data extraction were carried out independently by two investigators (A.M.P. and E.M.) according to the above-mentioned framework. Disagreement was resolved by discussion with a third investigator (F.B.). For the studies that fulfilled the criteria but presented data in aggregate so that the prevalence of the features of interest could not be estimated, the corresponding author was contacted. In case of no reply, the study was excluded. The following parameters were extracted: first author and year of publication, country of the study cohort, period of data collection, total number of cases fulfilling the criteria described in the Population of the PICOS framework, total number of cases with tumor necrosis, total number of cases with mitotic count ≥5 mitoses per 2 mm², type of study, type of tumors included in the study, and, when available, tumor subtypes.

The total number of cases fulfilling the criteria of DHGTC was obtained combining data on necrosis and mitosis. When overlapping cases could not be deducted, the number of DHGTC was defined as the minimum possible, thus with perfect overlapping between cases with necrosis and those with mitoses above the cutoff. When either data on necrosis or mitoses were provided, these were used to define the number of DHGTC. Data on the impact of necrosis and mitoses on prognosis were recorded when fulfilling the Population criteria of the PICOS. Quality of the studies included in the meta-analysis was assessed by a modification of the Newcastle–Ottawa scale for cohort and cross-sectional studies.⁹ Although this tool consists of 3 categories (i.e., Selection, Comparability, and Outcome) with a maximum of 9 items, Comparability and Outcome were considered only when data on prognosis were available. Hence, studies used to assess the prevalence of DHGTC could be awarded with a maximum of 4 stars (i.e., 4 items of Selection). The complete criteria are available on the Supplementary Data S2.

Statistics

A meta-analysis was conducted using the raw numbers extracted in the previous phase. The metaprop function of the meta R package v.6.2-1 was applied to estimate the prevalence of DHGTC within WDTC and in the subtypes for which at least two studies were collected. The univariate meta-analysis was carried out with a random-effect model using the inverse variance method and logit transformation. Continuity correction of 0.5 was applied to studies with zero cell frequencies, and confidence intervals of individual studies were estimated by the Clopper–Pearson method. Heterogeneity was assessed via the I ² and τ ² statistics.

The latter was computed with the DerSimonian–Laird estimator, and confidence interval was estimated by the Jackson method. Substantial heterogeneity was defined when I ² was above 50%, and p-value of Cochran's test was below 0.05,¹⁰ and the causes of such heterogeneity were further explored whenever present. Publication bias and small study effects were evaluated by funnel plot and Egger's test. In the presence of significant bias (p ≤ 0.1) and in the absence of substantial heterogeneity, the trim-and-fill method was used for adjusting the number and outcomes of missing studies. All analyses and graphics were produced in R environment (v. 4.2.2; www.r-project.org [last accessed November 13, 2023]).

Results

Literature search and quality assessment

Literature search identified 28 studies that met the criteria for inclusion; these were included in the systematic review and meta-analysis. The selection process is detailed in Figure 1. Twenty-four of the 28 studies included cases not selected nor enriched for patients' prognosis. These 24 studies (Table 2) were used to estimate the prevalence of DHGTC since the introduction of studies focusing on patients with poor prognosis might produce enrichment for aggressive pathological features, including necrosis and high mitotic count. Follicular thyroid carcinoma (FTC) represented about 6.1% of all cases. Four studies reported data on clinically selected patients (i.e., radioiodine-resistant tumors, metastatic PTC and FTC, fatal non-anaplastic tumors) and were used in a separate analysis (Table 3). Only four studies analyzed the prognostic impact of necrosis fulfilling the PICOS criteria. Sugitani et al. reported a significant association between necrosis and disease-specific survival, but the results were not confirmed on multivariate analysis.¹¹

FIG. 1.

PRISMA flow diagram. The detailed process that led to the inclusion of 25 studies in the meta-analysis is shown. For each step, raw numbers and reasons of exclusion are reported. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Table 2.

Studies Reporting Data on High-Grade Features Without Clinical Selection of Patients and Included in the Meta-Analysis

First author, year	Country	Data collection	Type of study	Study group	No. of cases ^a	Data on necrosis ^a	Data on mitosis ^a
Akslen, 2000²⁰	USA	1971–1985	Cohort	Consecutive PTC	110	Yes	No
Thompson, 2005³²	USA	1970–2000	Cross-sectional	Consecutive DSPTC	28	Yes	No
Arif, 2007²⁴	UK	NA	Cohort	Consecutive follicular thyroid lesions	76	Yes	No
Heikkilä, 2010³³	Finland	1990–2006	Cross-sectional	Consecutive follicular thyroid lesions	34	Yes	No
Sugitani, 2010¹¹	Japan	1994–2001	Cohort	Consecutive PTC and FTC	375	Yes	No
Chen, 2011³⁴	USA	1993–2005	Cross-sectional	COLPTC	9	Yes	No
Asioli, 2013³⁵	USA and Italy	1955–2004 (USA), 1992–2011 (Italy)	Cross-sectional	Consecutive PTC with hobnail features	24	Yes	Yes
Gnemmi, 2014³⁶	France	2000–2010	Cross-sectional	Consecutive well-differentiated tumors with solid, trabecular, or insular component	36	Yes	Yes
Lubitz, 2014³⁷	USA	2009–2012	Cross-sectional	Consecutive HPTC	10	Yes	No
Amacher, 2015³⁸	USA	2007–2011	Cross-sectional	Consecutive PTC with hobnail features	6	Yes	Yes
Chindris, 2015³⁹	USA	2001–2012	Cross-sectional	Consecutive oncocytic carcinoma	173	Yes^b	Yes^b
Lee, 2015⁴⁰	Korea	NA	Cross-sectional	Consecutive HPTC	10	Yes	Yes
Tallini, 2015⁴¹	Italy	1994–2011	Cohort	Consecutive PTC ≤1 cm	296	Yes	Yes
Ieni, 2016⁴²	Italy	2010–2015	Cross-sectional	Consecutive HPTC	8	Yes	Yes
Morandi, 2017⁴³	Italy	NA	Cross-sectional	HPTC	18	Yes	Yes
Özçelik, 2019⁴⁴	Turkey	2000–2012	Cohort	PTC	197	Yes	No
Chu, 2020⁴⁵	USA	2013–2020	Cross-sectional	Fusion-positive well-differentiated tumors	57	Yes	Yes
Cracolici, 2020²⁵	USA	1987–2018	Cross-sectional	Consecutive FTC	35	Yes	No
Póvoa, 2020¹²	Portugal	2002–2015	Cohort	Consecutive PTC	237	Yes	No
Suster, 2020⁴⁶	USA	NA	Cross-sectional	PTC with myofibroblastic stromal component	7	Yes	Yes
Vural, 2021⁴⁷	Turkey	2010–2020	Cross-sectional	SPTC	28	Yes	Yes
Saliba, 2022¹⁵	USA	1994–2018	Cohort	Follicular cell-derived tumors in patients aged 21 years or younger	209	Yes	Yes
Chou, 2023¹⁶	Australia	2000–2022	Cross-sectional	Consecutive DSPTC	43	No	Yes
Jeong, 2023¹⁴	Korea	2019–2020	Cohort	Consecutive DTC	1069	Yes	Yes

Cases fulfilling the inclusion criteria and for which raw numbers were available.

Data on necrosis and mitosis provided in aggregate.

COLPTC, columnar subtype papillary thyroid carcinoma; DSPTC, diffuse sclerosing subtype papillary thyroid carcinoma; FTC, follicular thyroid carcinoma; HPTC, hobnail subtype papillary thyroid carcinoma; PTC, papillary thyroid carcinoma; SPTC, solid subtype papillary thyroid carcinoma.

Table 3.

Studies Reporting Data on High-Grade Features in Clinically Selected Patients

First author, year	Country	Data collection	Type of study	Study group	No. of cases ^a	Data on necrosis ^a	Data on mitosis ^a
Rivera, 2008¹⁷	USA	1996–2003	Cross-sectional	RAIR tumors	35	Yes^b	Yes^b
Deandreis, 2011¹⁸	France	2002–2007	Cross-sectional	Metastatic PTC and FTC	56	Yes	No
Xu, 2016¹⁹	USA	1985–2010	Cross-sectional	Fatal non-anaplastic tumors	43	Yes^b	Yes^b
Metovic, 2023¹³	Italy	2002–2017	Cross-sectional	Clinically aggressive PTC^c	43	Yes	Yes

Cases fulfilling the inclusion criteria and for which raw numbers were available.

Data on necrosis and mitosis provided on aggregate.

Presence of metastases at diagnosis, development of distant metastases during follow-up, biochemical recurrence defined as >2 ng/mL levels of thyroglobulin after thyrotropin stimulation but in the absence of structural disease.

RAIR, radioiodine resistant.

Similar results were obtained by Póvoa et al., which observed the association between necrosis and recurrent or persistent disease only at univariate analysis.¹² In line with these findings, in the study by Metovic et al.,¹³ necrosis and mitotic count were associated with disease-free and overall survival at univariate analysis but were not independent predictors after adjusting for confounders. Finally, necrosis and high mitotic count were significant predictors of recurrence in the study by Jeong et al.,¹⁴ but only necrosis remained significant at multivariate analysis. These data could not be used for meta-analytic purposes due to the heterogeneity of endpoints.

In all studies, the data derive from secure records (i.e., surgical records or histological revision). The overall quality was superior for cohort studies in comparison with cross-sectional ones. The complete quality assessment of the studies is reported in Supplementary Table S2.

Meta-analysis of DHGTC prevalence in clinically unselected patients

The prevalence of DHGTC was assessed on 3095 WDTC from 24 studies without prior clinical selection. Raw numbers of tumor subtypes in this cohort are reported in Figure 2. Considering the entire WDTC cohort, 138 tumors could be reclassified as DHGTC. The random-effect model resulted in a pooled prevalence of 0.072 [95% confidence interval, CI, = 0.045–0.113] with a substantial heterogeneity among studies (I ² = 84% [CI = 77–88%], p < 0.0001; Supplementary Fig. S1A). Hence, we began by stratifying cases by tumor growth pattern, which might account for such heterogeneity. The prevalence of DHGTC in all PTC (n = 2696) was 0.059 [CI = 0.036–0.097], again with substantial but lower heterogeneity (I ² = 80% [CI = 69–87%], p < 0.0001; Supplementary Fig. S1B).

FIG. 2.

Cases without prior clinical selection included in the meta-analysis. The diagram details all the tumor subtypes. *Other PTC, not specified by the authors; other than CPTC, FPTC, TCPTC, and STPTC. COLPTC, columnar subtype papillary thyroid carcinoma; CPTC, classic subtype papillary thyroid carcinoma; DSPTC, diffuse sclerosing subtype papillary thyroid carcinoma; FPTC, follicular subtype papillary thyroid carcinoma; FTC, follicular thyroid carcinoma; HPTC, hobnail subtype papillary thyroid carcinoma; MIFTC, minimally invasive follicular thyroid carcinoma; NOS, not otherwise specified; OCA, oncocytic carcinoma; PTC, papillary thyroid carcinoma; STPTC, solid trabecular subtype papillary thyroid carcinoma; TCPTC, tall cell subtype papillary thyroid carcinoma; WDTC, well-differentiated thyroid carcinoma; WIFTC, widely invasive follicular thyroid carcinoma.

On the contrary, the estimated proportion of high-grade cases in FTC (n = 189) was greater than that in PTC: 0.146 [CI = 0.101–0.205] with I ² = 0% ([CI = 0–75%], p = 0.74; Supplementary Fig. S1C). Then, a further stratification was performed based on PTC subtypes. The lowest proportion of tumors reclassified as high-grade was observed in diffuse sclerosing subtype papillary thyroid carcinoma (DSPTC, n = 82) after removing pediatric patients (i.e., the study by Saliba et al.¹⁵ and 10 cases from the study by Chou et al.¹⁶) that would have introduced significant heterogeneity: 0.018 [CI = 0.004–0.084], I ² = 0% ([CI = 0–90%], p = 0.98; Supplementary Fig. S2A). The proportion of classic subtype papillary thyroid carcinoma (CPTC, n = 494) that could be reclassified as high-grade was 0.042 [CI = 0.027–0.066] with I ² = 0% ([CI = 0–71%], p = 0.49; Supplementary Fig. S2B). In follicular subtype papillary thyroid carcinoma (FPTC, n = 216), the random-effect model resulted in a proportion of DHGTC of 0.076 [CI = 0.013–0.341], but the amount of heterogeneity was significant (I ² = 79.1% [CI = 54.3–90.5%], p = 0.0002).

The major source of heterogeneity was attributed to the pediatric cohort from the study by Saliba et al. Hence, we removed this study and re-ran the analysis (n = 196) that resulted in a pooled proportion of 0.036 ([CI = 0.010–0.124]; I ² = 31%, [CI = 0–73%], p = 0.22; Supplementary Fig. S2C). The proportion of high-grade tumors was higher in solid trabecular subtype papillary thyroid carcinoma (STPTC, n = 57): 0.154 [CI = 0.067–0.314], I ² = 17% ([CI = 0–83%], p = 0.31; Supplementary Fig. S2D), as well as in tall cell subtype papillary thyroid carcinoma (TCPTC, n = 89): 0.187 [CI = 0.064–0.437], I ² = 54% ([CI = 0–83%], p = 0.07; Supplementary Fig. S2E).

After adjusting for potential publication bias in the absence of significant heterogeneity, the estimated proportion of DHGTC in the hobnail subtype papillary thyroid carcinoma (HPTC, n = 58) raised to 0.202 [CI = 0.114–0.334], I ² = 0% ([CI = 0–71%], p = 0.54; Supplementary Fig. S2F). Finally, we computed the cumulative proportion of high-grade tumors in those histological PTC subtypes considered to be clinically aggressive (i.e., TCPTC, HPTC, and columnar subtype papillary thyroid carcinoma⁵ [COLPTC], n = 165): 0.168 [CI = 0.108–0.252], I ² = 9% ([CI = 0–48%], p = 0.35; Supplementary Fig. S2G).

Similar to PTC, FTC was further stratified into minimally invasive follicular thyroid carcinoma (MIFTC) and widely invasive follicular thyroid carcinoma (WIFTC). In clinically unselected patients, the proportion of high-grade tumors was similar in MIFTC (n = 37, 0.136 [CI = 0.058–0.287], I ² = 0% [CI = 0–0%], p = 0.81; Supplementary Fig. S3A) and WIFTC (n = 73, 0.152 [CI = 0.086–0.254], I ² = 0% [CI = 0–0%], p = 0.60; Supplementary Fig. S3B). Significant publication bias was detected in HPTC only (p = 0.04), and estimates were adjusted as mentioned previously. All funnel plots are reported in Supplementary Figure S4A–I and L. The summary of the results in clinically unselected patients is reported in Figure 3.

FIG. 3.

Summary of results. The graph shows the prevalence (diamonds) and relative confidence interval (black lines) in clinically unselected cases and by tumor subtypes along with the total number of cases, the total number of studies, and heterogeneity.

Prevalence of high-grade tumors in clinically selected well-differentiated tumors

Since the impact of high-grade features on prognosis could not be estimated due to the lack of studies with common endpoints that met the inclusion criteria for the meta-analysis, we tried to assess it indirectly. In four studies^13,17
–19 (Table 3 and Supplementary Fig. S5), the presence of tumor necrosis and/or mitotic count were assessed in cohorts of patients with poor prognosis (i.e., radioiodine-resistant tumors, metastatic PTC and FTC, fatal non-anaplastic tumors, cases with biochemical recurrence). In these cases (n = 177), the pooled proportion of DHGTC was 0.287 [CI = 0.155–0.469], I ² = 74% ([CI = 28–91%], p = 0.008; Supplementary Fig. S6). The main cause of such heterogeneity was attributed to the study by Metovic et al., in which cases with biochemical recurrence but no evidence of structural disease was included. Without this study, heterogeneity (I ²) dropped to 9% ([CI = 0–91%], p = 0.33). No significant publication bias was detected (p = 0.24; Supplementary Fig. S7).

Discussion

Grading of follicular cell-derived tumors, originally proposed in the year 2000,^2,20 has been introduced in the current fifth edition of the WHO classification of thyroid neoplasms. The current classification scheme distinguishes two types of high-grade tumors: PDTC and DHGTC. The latter represents a new entry, which is characterized by the presence of tumor necrosis and/or high mitotic count (i.e., ≥5 per 2 mm²) but not fulfilling the Turin criteria for PDTC. Since high-grade features have been often presented in aggregate data with PDTC and, sometimes, anaplastic tumors, the impact of this new DHGTC entry in the classification scheme is yet to be determined. In this study, we performed a systematic review and meta-analysis aimed at estimating the prevalence of high-grade tumors within WDTC and per tumor subtype. Pooled estimates were performed from studies reporting clinically unselected patients. Most of these studies presented consecutive cases series. The overall proportion of FTC (i.e., about 6.1%) is in line with previous literature reports.^3,21

Similarly, the cumulative rate (i.e., between 6.1% and 15.2%, computed on the total of PTC and without PTC not otherwise specified, respectively) of aggressive PTC subtypes (i.e., TCPTC, COLPTC, and HPTC) is consistent with that of consecutive series.²² The aggregate prevalence of DHGTC in WDTC is about 7.2% of all cases. In accordance to the large clinical–pathological variability across tumor subtypes, heterogeneity among studies was high. To reduce heterogeneity, a first stratification was performed based on tumor growth pattern. The prevalence of high-grade features in FTC was more than double than that in PTC (14.6% vs. 5.9%). This result is in line with the poorer overall prognosis of FTC.²³ However, three considerations should be made. First, in this meta-analysis, WIFTC was more abundant than MIFTC (i.e., 73 vs. 37), which is not common in real-life series. Second, although the prevalence of high-grade tumors in MIFTC was only slightly lower than that in WIFTC (i.e., 13.6% vs. 15.2%), the former might encompass both true MIFTC and angioinvasive FTC.

Finally, in the case series described by Arif et al. and Cracolici et al.,^24,25 some FTC show solid, trabecular, or insular areas; hence, the possibility that a part of them could be reclassified as PDTC cannot be excluded. As a consequence, the rate of high-grade FTC here reported might be overestimated, thus highlighting that, in FTC, the distinction between high grade and poorly differentiated might be subjected to interobserver variability. In regard to PTC subtypes, DSPTC showed the lowest rate of high-grade features (i.e., 1.8%) with the exception of tumors of pediatric age. Indeed, in the cohort described by Saliba et al.,¹⁵ 10 of 32 (31%) pediatric DSPTC could be reclassified as DHGTC. As expected, a low proportion of FPTC presented high-grade features (i.e., 3.5%)—again with the exception of pediatric tumors¹⁵—followed by CPTC (4.2%). A remarkably higher prevalence of DHGTC was observed in STPTC (15.3%) and in those histological PTC subtypes considered to be clinically aggressive (i.e., TCPTC, COLPTC, and HPTC, 16.8%).

While these findings are consistent with the aggressive clinical–pathological behavior of TCPTC, COLPTC, and HPTC,²² controversies exist around STPTC. In fact, in a meta-analysis, it was demonstrated that patients with solid PTC have a higher risk of recurrence and mortality than those with CPTC,²⁶ and some authors place STPTC among aggressive PTC subtypes. On the contrary, a recent study reported an excellent prognosis of patients with STPTC, especially those with encapsulated noninvasive tumors.²⁷ This controversy may be related to the lack of consensus around the proportion of tumor cells showing solid/trabecular growth pattern to define the tumor as STPTC as well as to the fact that at least 3 of the 11 publications reviewed in the meta-analysis of Vuong et al.²⁶ explicitly included cases with elevated mitotic count^28,29 or necrosis.²⁰ According to the current WHO indications, the diagnosis of solid PTC subtype requires that >50% of tumor show a solid, trabecular, or nested growth pattern.³

Due to the lack of studies that met the inclusion criteria and with the same endpoint, to assess the prognostic impact of high-grade features on the prognosis of patients with WDTC, we estimated the proportion of cases that could be reclassified as DHGTC in four series of clinically selected patients (i.e., patients with biochemical recurrence, patients with radioiodine resistant tumors, patients with distant metastases, and patients who died of disease^17
–19). Almost one third (i.e., 28.7%) of these tumors presented high-grade features. It has to be acknowledged that, while the proportion of FTC in this cohort was similar to that of the clinically unselected one (7.9% vs. 6.1%), aggressive PTC accounted for about 30% of all PTC. Nevertheless, the pooled proportion of DHGTC in patients with poor prognosis was much higher than that in clinically unselected aggressive PTC (i.e., 16.8%), thus supporting a negative impact of high-grade features on the prognosis of patients with WDTC.

This study has some limitations. First, the estimates reported here derive from studies from various countries, different design, and with potential interobserver variability. Also, the estimates are mainly based on case series reporting the assessment of the features which define DHGTC (i.e., necrosis and mitotic count) and do not represent histological revisions specifically intended to assess DHGTC. The sample size of a few studies is small, thus suboptimal to estimate the prevalence of DHGTC. However, this is a common situation for reports of rare tumor types and subtypes. It has to be acknowledged that our estimates are conservative. In fact, a perfect overlap between the presence of necrosis and mitotic count above the cutoff was considered whenever this information was missing. Also, some studies did not evaluate mitotic counts, and the prevalence of DHGTC was solely based on the presence of necrosis. Due to the lack of assessment of mitotic counts in some studies, a higher prevalence of DHGTC especially among aggressive subtypes cannot be excluded.

Nevertheless, the estimates provided in this study were mostly generated using consecutive series and in controlled or absence of substantial heterogeneity; hence, the prevalence here provided can be considered representative of real-life data. The case of MIFTC may be an exception. Indeed, in this study, we could not distinguish angioinvasive tumors from true MIFTC (i.e., tumors with capsular invasion only). Regarding FPTC, the current edition of the WHO system has introduced several novelties, that is, the distinction between the invasive encapsulated FPTC and the infiltrative FPTC subtype;³ in the present study, all FPTC were considered as a single group.

Also, the hobnail subtype was acknowledged in the WHO classification scheme only in 2017; hence, some cases might not have been diagnosed in earlier reports. Nevertheless, the criteria for the diagnosis of the hobnail subtypes were first outlined in 2004³⁰ and further refined in 2009.³¹ Therefore, we do not expect substantial underestimation of the hobnail subtype. Finally, the association of high-grade features with poor prognosis could only be inferred indirectly, due to the lack of studies that met the inclusion criteria for the meta-analysis.

In conclusion, the introduction of DHGTC in the latest edition of the WHO may lead to the histological reclassification of a number of tumors of follicular cell derivation that is not negligible. The proportion of high-grade features is low in DSPTC, FPTC and CPTC, while in STPTC and histologically aggressive PTC subtypes, it can raise up to 20%. Significant differences could be observed in tumors of pediatric age, among which DSPTC and FPTC may frequently present with necrosis and/or high mitotic index. The reclassification of tumors as DHGTC will likely improve the prognostic stratification of patients with WDTC. This new scheme will identify patients who need aggressive therapeutic approach and more intensive follow-up. Also, this change will probably determine an enrichment of BRAF-like tumors (i.e., PTC, especially aggressive subtypes) within high-grade follicular cell-derived carcinomas. Dedicated studies should further assess the impact of high-grade features on prognosis of patients with specific WDTC subtypes.

Footnotes

Acknowledgments

We are grateful to Prof. Iwao Sugitani from the Nippon Medical School, Tokyo, Japan, Prof. Sofia Asioli from the University of Bologna, Italy, and Prof. Marco Volante from the University of Turin, Italy, for providing additional data and helpful information.

Authors' Contributions

Conceptualization: A.M.P. and F.B. Methodology: A.M.P., E.M., R.A.G., G.T., and F.B. Formal analysis: A.M.P. Writing—original draft: A.M.P. Writing—review and editing: all authors. Supervision: F.B.

Author Disclosure Statement

All authors have no disclosure to declare.

Funding Information

This study was supported in part by the Cancer Center Support Grant of the National Institutes of Health/National Cancer Institute under award number P30CA008748.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Figure S6

Supplementary Figure S7

Supplementary Table S1

Supplementary Table S2

Supplementary Data S1

Supplementary Data S2

References

Volante

, Collini

, Nikiforov

, et al. Poorly differentiated thyroid carcinoma: The Turin proposal for the use of uniform diagnostic criteria and an algorithmic diagnostic approach. Am J Surg Pathol, 2007; 31(8):1256–1264; doi: 10.1097/PAS.0b013e3180309e6a

Tallini

Poorly differentiated thyroid carcinoma. Are we there yet?. Endocr Pathol, 2011; 22(4):190–194; doi: 10.1007/s12022-011-9176-5

WHO Classification of Tumors Editorial Board: Endocrine and Neuroendocrine Tumors. 5th edition. WHO Classification of Tumours Series. International Agency for Research on Cancer: Lyon, France; 2022.

Basolo

, Macerola

, Poma

, et al. The 5th edition of WHO classification of tumors of endocrine organs: Changes in the diagnosis of follicular-derived thyroid carcinoma. Endocrine, 2023; 80(3):470–476; doi: 10.1007/s12020-023-03336-4

Baloch

, Asa

, Barletta

, et al. Overview of the 2022 WHO classification of thyroid neoplasms. Endocr Pathol, 2022; 33(1):27–63; doi: 10.1007/s12022-022-09707-3

Skansing

, Londero

, Asschenfeldt

, et al. Nonanaplastic follicular cell-derived thyroid carcinoma: Mitosis and necrosis in long-term follow-up. Eur Arch Otorhinolaryngol, 2017; 274(6):2541–2548; doi: 10.1007/s00405-017-4527-6

Wong

, Dong

, Telatar

, et al. Papillary thyroid carcinoma with high-grade features versus poorly differentiated thyroid carcinoma: An analysis of clinicopathologic and molecular features and outcome. Thyroid, 2021; 31(6):933–940; doi: 10.1089/thy.2020.0668

Page

, McKenzie

, Bossuyt

, et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 2021; 372:n71; doi: 10.1136/bmj.n71

Wells

, Shea

, O'Connell

, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. n.d.

10.

Higgins

JPT

, Thomas

, Chandler

, et al. Cochrane Handbook for Systematic Reviews of Interventions Version 6.3. 2022.

11.

Sugitani

, Toda

, Yamamoto

, et al. Re-evaluation of histopathological factors affecting prognosis of differentiated thyroid carcinoma in an iodine-sufficient country. World J Surg, 2010; 34(6):1265–1273; doi: 10.1007/s00268-009-0305-y

12.

Póvoa

, Teixeira

, Bella-Cueto

, et al. Clinicopathological features as prognostic predictors of poor outcome in papillary thyroid carcinoma. Cancers, 2020; 12(11):1–16; doi: 10.3390/cancers12113186

13.

Metovic

, Cabutti

, Osella-Abate

, et al. Clinical and pathological features and gene expression profiles of clinically aggressive papillary thyroid carcinomas. Endocr Pathol, 2023; 34(3):298–310; doi: 10.1007/s12022-023-09769-x

14.

Jeong

, Kim

, Yu

, et al. Incidence and clinicopathological features of differentiated high-grade thyroid carcinomas: An institutional experience. Endocr Pathol, 2023; 34(3):287–297; doi: 10.1007/s12022-023-09778-w

15.

Saliba

, Alzumaili

, Katabi

, et al. Clinicopathologic and prognostic features of pediatric follicular cell-derived thyroid carcinomas: A retrospective study of 222 patients. Am J Surg Pathol, 2022; 46(12):1659–1669; doi: 10.1097/PAS.0000000000001958

16.

Chou

, Qiu

, Crayton

, et al. A detailed histologic and molecular assessment of the diffuse sclerosing variant of papillary thyroid carcinoma. Mod Pathol, 2023; 36(12):100329; doi: 10.1016/j.modpat.2023.100329

17.

Rivera

, Ghossein

, Schoder

, et al. Histopathologic characterization of radioactive iodine-refractory fluorodeoxyglucose-positron emission tomography-positive thyroid carcinoma. Cancer, 2008; 113(1):48–56; doi: 10.1002/cncr.23515

18.

Deandreis

, Al Ghuzlan

, Leboulleux

, et al. Do histological, immunohistochemical, and metabolic (radioiodine and fluorodeoxyglucose uptakes) patterns of metastatic thyroid cancer correlate with patient outcome?. Endocr Relat Cancer, 2011; 18(1):159–169; doi: 10.1677/ERC-10-0233

19.

, Ibrahimpasic

, Wang

, et al. Clinicopathologic features of fatal non-anaplastic follicular cell-derived thyroid carcinomas. Thyroid, 2016; 26(11):1588–1597; doi: 10.1089/thy.2016.0247

20.

Akslen

, LiVolsi

. Prognostic significance of histologic grading compared with subclassification of papillary thyroid carcinoma. Cancer, 2000; 88(8):1902–1908.

21.

Grani

, Lamartina

, Durante

, et al. Follicular thyroid cancer and Hürthle cell carcinoma: Challenges in diagnosis, treatment, and clinical management. Lancet Diabetes Endocrinol, 2018; 6(6):500–514; doi: 10.1016/S2213-8587(17)30325-X

22.

Baloch

, LiVolsi

. Special types of thyroid carcinoma. Histopathology, 2018; 72(1):40–52; doi: 10.1111/his.13348

23.

Liu

, Zeng

, Huang

, et al. Prognosis of FTC compared to PTC and FVPTC: Findings based on SEER database using propensity score matching analysis. Am J Cancer Res, 2018; 8(8):1440–1448.

24.

Arif

, Patel

, Blanes

, et al. Cytoarchitectural and kinetic features in the histological evaluation of follicular thyroid neoplasms. Histopathology, 2007; 50(6):750–763; doi: 10.1111/j.1365-2559.2007.02680.x

25.

Cracolici

, Ritterhouse

, Segal

, et al. Follicular thyroid neoplasms: Comparison of clinicopathologic and molecular features of atypical adenomas and follicular thyroid carcinomas. Am J Surg Pathol, 2020; 44(7):881–892; doi: 10.1097/PAS.0000000000001489

26.

Vuong

, Odate

, Duong

UNP

, et al. Prognostic importance of solid variant papillary thyroid carcinoma: A systematic review and meta-analysis. Head Neck, 2018; 40(7):1588–1597; doi: 10.1002/hed.25123

27.

, Viswanathan

, Zhang

, et al. The solid variant of papillary thyroid carcinoma: A multi-institutional retrospective study. Histopathology, 2022; 81(2):171–182; doi: 10.1111/his.14668

28.

Nikiforov

, Erickson

, Nikiforova

, et al. Solid variant of papillary thyroid carcinoma: incidence, clinical-pathologic characteristics, molecular analysis, and biologic behavior. Am J Surg Pathol, 2001; 25(12):1478–1484; doi: 10.1097/00000478-200112000-00002

29.

Nikiforov

, Gnepp

. Pediatric thyroid cancer after the Chernobyl disaster. Pathomorphologic study of 84 cases (1991–1992) from the Republic of Belarus. Cancer, 1994; 74(2):748–766; doi: 10.1002/1097-0142(19940715)74:2<748::aid-cncr2820740231>3.0.co;2-h

30.

Kakudo

, Tang

, Ito

, et al. Papillary carcinoma of the thyroid in Japan: Subclassification of common type and identification of low risk group. J Clin Pathol, 2004; 57(10):1041–1046; doi: 10.1136/jcp.2004.017889

31.

Motosugi

, Murata

S-I

, Nagata

, et al. Thyroid papillary carcinoma with micropapillary and hobnail growth pattern: A histological variant with intermediate malignancy?. Thyroid, 2009; 19(5):535–537; doi: 10.1089/thy.2008.0271

32.

Thompson

LDR

, Wieneke

, Heffess

. Diffuse sclerosing variant of papillary thyroid carcinoma: A clinicopathologic and immunophenotypic analysis of 22 cases. Endocr Pathol, 2005; 16(4):331–348; doi: 10.1385/ep:16:4:331

33.

Heikkilä

, Siironen

, Hagström

, et al. Follicular thyroid neoplasm: Clinicopathologic features suggesting malignancy. APMIS, 2010; 118(11):846–854; doi: 10.1111/j.1600-0463.2010.02668.x

34.

Chen

J-H

, Faquin

, Lloyd

, et al. Clinicopathological and molecular characterization of nine cases of columnar cell variant of papillary thyroid carcinoma. Mod Pathol, 2011; 24(5):739–749; doi: 10.1038/modpathol.2011.2

35.

Asioli

, Erickson

, Righi

, et al. Papillary thyroid carcinoma with hobnail features: Histopathologic criteria to predict aggressive behavior. Hum Pathol, 2013; 44(3):320–328; doi: 10.1016/j.humpath.2012.06.003

36.

Gnemmi

, Renaud

, Do Cao

, et al. Poorly differentiated thyroid carcinomas: Application of the Turin proposal provides prognostic results similar to those from the assessment of high-grade features. Histopathology, 2014; 64(2):263–273; doi: 10.1111/his.12246

37.

Lubitz

, Economopoulos

, Pawlak

, et al. Hobnail variant of papillary thyroid carcinoma: An institutional case series and molecular profile. Thyroid, 2014; 24(6):958–965; doi: 10.1089/thy.2013.0573

38.

Amacher

, Goyal

, Lewis

, et al. Prevalence of a hobnail pattern in papillary, poorly differentiated, and anaplastic thyroid carcinoma: A possible manifestation of high-grade transformation. Am J Surg Pathol, 2015; 39(2):260–265; doi: 10.1097/PAS.0000000000000329

39.

Chindris

A-M

, Casler

, Bernet

, et al. Clinical and molecular features of Hürthle cell carcinoma of the thyroid. J Clin Endocrinol Metab, 2015; 100(1):55–62; doi: 10.1210/jc.2014-1634

40.

Lee

, Kim

, Jeon

, et al. Cytologic, clinicopathologic, and molecular features of papillary thyroid carcinoma with prominent hobnail features: 10 Case reports and systematic literature review. Int J Clin Exp Pathol, 2015; 8(7):7988–7997.

41.

Tallini

, de Biase

, Durante

, et al. BRAF V600E and risk stratification of thyroid microcarcinoma: A multicenter pathological and clinical study. Mod Pathol, 2015; 28(10):1343–1359; doi: 10.1038/modpathol.2015.92

42.

Ieni

, Barresi

, Cardia

, et al. The micropapillary/hobnail variant of papillary thyroid carcinoma: A review of series described in the literature compared to a series from one southern Italy pathology institution. Rev Endocr Metab Disord, 2016; 17(4):521–527; doi: 10.1007/s11154-016-9398-4

43.

Morandi

, Righi

, Maletta

, et al. Somatic mutation profiling of hobnail variant of papillary thyroid carcinoma. Endocr Relat Cancer, 2017; 24(2):107–117; doi: 10.1530/ERC-16-0546

44.

Özçelik

, Bircan

, Sarıkaya Ş, et al. BRAF V600E mutation in papillary thyroid cancer is correlated with adverse clinicopathological features but not with iodine exposure. Endokrynol Pol, 2019; 70(5):401–408; doi: 10.5603/EP.a2019.0025

45.

Chu

Y-H

, Wirth

, Farahani

, et al. Clinicopathologic features of kinase fusion-related thyroid carcinomas: An integrative analysis with molecular characterization. Mod Pathol, 2020; 33(12):2458–2472; doi: 10.1038/s41379-020-0638-5

46.

Suster

, Michal

, Nishino

, et al. Papillary thyroid carcinoma with prominent myofibroblastic stromal component: Clinicopathologic, immunohistochemical and next-generation sequencing study of seven cases. Mod Pathol, 2020; 33(9):1702–1711; doi: 10.1038/s41379-020-0539-7

47.

Vural Ç, Kiraz

, Turan

, et al. Solid variant of papillary thyroid carcinoma: An analysis of 28 cases with current literature. Ann Diagn Pathol, 2021; 52:151737; doi: 10.1016/j.anndiagpath.2021.151737