Abstract
Thyroid tumors with questionable capsular invasion (CI) or angioinvasion (AI) pose diagnostic and management challenges. The World Health Organization (WHO) recognizes these borderline entities as follicular tumors of uncertain malignant potential (FT-UMP) or well-differentiated tumors of uncertain malignant potential (WDT-UMP). We have conducted a retrospective, single-center study of all thyroid tumors diagnosed as UMP (2005–2025). Of 534 tumors initially labeled “uncertain malignant potential,” application of exclusion criteria yielded 324 tumors from 318 patients with questionable CI and/or AI. Clinicopathological characteristics, molecular testing data, postoperative laboratory values, imaging, and outcomes were analyzed. UMP was diagnosed in 0.5% of thyroid surgeries; median age 53-years and 2:1 female predominance. Cohort classification included oncocytic FT-UMP (OFT-UMP) (40%), FT-UMP (34%) and WDT-UMP (26%). Median tumor size was 3.2 cm. Molecular testing yielded RAS variants most frequently (79%), followed by RAS-like variants (4%), gene fusions (4%), PTEN loss (4%), and TERT promoter alterations (3%). With follow-up extending up to 15 years, no recurrences or metastases were recorded, supporting the UMP designation as tumors of low biological potential and favoring conservative management. Longer-term, multi-institutional studies will refine risk stratification and clinical management.
Keywords: Thyroid, FT-UMP, WDT-UMP, OFT-UMP, Uncertain malignant potential, Molecular typing, RAS
Introduction
A century ago, Graham noted the designation of thyroid tumors as “benign” versus “carcinoma” was problematic, given some carcinomas behaved indolently, and rarely, benign-appearing tumors metastasized [1]. Tumors not fitting into this traditional binary scheme complicated simple classification. Prior to the description of papillary thyroid carcinoma (PTC)-like nuclear atypia in 1960 [2], all non-invasive encapsulated tumors without papillary architecture were classified as follicular thyroid adenomas, with invasive tumors classified as follicular thyroid carcinomas [3]. In 1954, Hazard and Kenyon described follicular thyroid adenomas with increased proliferation and unusual growth patterns [4]. Over the next two decades, further refinements produced the term atypical thyroid adenoma [5–7].
In 1980, Lang et al. investigated atypical adenomas, highlighting the importance of examining hematoxylin and eosin (H&E) tissue levels to detect capsular invasion (CI) and angioinvasion (AI). They concluded that AI is a more reliable indicator of likely malignancy [3]. Updated criteria for invasion of tumor capsule and blood vessels were formalized by Franssila et al. in the article on follicular carcinoma published in the Proceedings of the 1984 International Workshop on Thyroid Tumor Pathology [8]. Williams and the Chernobyl Pathologists Group (2000) proposed a pragmatic three-part nomenclature introducing the concept of tumors of uncertain malignant potential (UMP). The term well-differentiated tumor of uncertain malignant potential (WDT-UMP) was applied to encapsulated tumors with PTC-like nuclear atypia, no AI and absent or questionable CI. Second, well-differentiated carcinoma not otherwise specified (WDC-NOS) captured encapsulated tumors with questionable PTC-like nuclear atypia with obvious invasion of tumor capsule and/or blood vessels. Third, follicular tumor of uncertain malignant potential (FT-UMP) described tumors lacking nuclear atypia or AI but with foci questionable for CI. Questionable foci were defined by an irregular tumor capsule/parenchymal interface or tumor cells entrapped by fibrosis within the capsule [9]. Follow-up studies highlighted the difficulty of diagnosing atypical adenomas and the challenges associated with diagnostic uncertainty [10–13]. Subsequent clinicopathologic and molecular studies have documented that these borderline entities are uncommon following extensive work-up. They may harbor RAS or RAS-like molecular alterations and are largely understood to have indolent but uncertain biological behavior. [14–16].
The reclassification of noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) in 2016 and its incorporation into the 2017 World Health Organization (WHO) Classification of Endocrine Tumors, formally recognized the “borderline/uncertain malignant potential” category for selected encapsulated follicular-patterned tumors, including FT-UMP and WDT-UMP [17, 18]. Most recently, the 5th edition of the WHO classification (2022) emphasized rigorous assessment for capsular and vascular invasion, specifically, angioinvasion. When foci of capsular invasion or angioinvasion are questionable, tumors with papillary thyroid carcinoma (PTC)-like nuclei are classified as well-differentiated tumors of uncertain malignant potential (WDT-UMP), and tumors with typical follicular nuclear morphology are classified as follicular tumors of uncertain malignant potential (FT-UMP) [19]. To our knowledge, this is the first and largest study to focus specifically on a pure UMP cohort, targeting the clinicopathological and molecular features and describing the clinical course of patients whose tumors were initially diagnosed with this designation.
Materials and Methods
Study Population
This study was approved by the Mass General Brigham Institutional Review Board (2011P000013 to PMS). The flowchart of UMP cohort selection is illustrated in Fig. 1. The Massachusetts General Hospital Laboratory Information Systems database (CoPath Plus; Sunquest, Tucson, AZ) was searched to identify all patients diagnosed with thyroid tumor of uncertain malignant potential (UMP) between 2005 and mid-2025 at Massachusetts General Hospital or Massachusetts Eye and Ear. A total of 527 cases (534 tumors) were initially identified by querying pathology reports for the term “uncertain malignant potential” appearing in the diagnosis, comment or descriptive microscopic examination. Of these, 137 tumors (from 136 patients) were subsequently excluded because the original pathologist favored a benign lesion, including follicular adenomas or adenomatous nodules (15 tumors), or because the descriptive microscopic findings were consistent with non-invasive follicular thyroid neoplasms with papillary-like nuclear features (NIFTP) diagnosis according to the most recent WHO classification (122 tumors). In addition, 34 cases (34 tumors) were excluded when the original pathologist favored carcinoma based on the presence of CI or AI. The final UMP cohort included 318 patients (342 tumors) with questionable capsular invasion (Fig. 2A, B) and/or questionable angioinvasion (Fig. 2C, D). To strictly adhere to the WHO classification of UMP, well-circumscribed tumors with atypical histologic features apart from questionable capsular/vascular invasion were also excluded from this study (39 patients; 39 tumors). To maintain a pure dataset and reflect routine diagnostic practice in a specialized tertiary care practice, cases were included based on data from the original pathology reports and cases were not microscopically re-evaluated. All patients were initially diagnosed and treated as UMP. Pathologists (BAA and PMS) were blinded to patient outcomes.
Fig. 1.
Flowchart of UMP cohort selection: From 527 patients (534 tumors), cases were excluded if the original pathologist favored carcinoma (based on CI or AI) or benign/NIFTP diagnoses. Tumors with questionable CI or AI constituted the UMP cohort (318 patients). Tumors with a well-circumscribed capsule but with atypical features (39 patients) were excluded from the study. *Atypical features included either: (1) tumors with 30–50% solid/trabecular growth and PTC-like nuclear atypia; (2) tumors with > 50% solid/trabecular growth lacking PTC-like nuclear atypia but showing 3–5 mitoses per 2 mm2; or (3) tumors showing procedural artifacts (e.g., capsular disruption or diffuse fine-needle aspiration–related changes such as infarction or squamous metaplasia). PTC, papillary thyroid carcinoma; CI, capsular invasion; AI, angioinvasion; NIFTP: noninvasive follicular thyroid neoplasm with papillary-like nuclear features
Fig. 2.
Morphologic features of two FT-UMP, one with questionable foci of capsular invasion (A, B) and one with a questionable focus of angioinvasion (C, D). Whole slide image of FT-UMP with central area of fibrosis, hyalinization, peripheral irregularities and questionable foci of capsular invasion (A; H&E, WSI). The questionable foci are pushing the periphery of the tumor into the adjacent normal thyroid (B; H&E, 100X), but the interface between the questionable foci and the adjacent normal thyroid shows well-delineation and minimal fibrosis (B-inset; H&E, 200X). FT-UMP with questionable focus for angioinvasion (C; H&E, 200X) which was no longer present on the deeper H&E level (D; H&E, 200X). FT-UMP: follicular tumor of uncertain malignant potential, H&E: Hematoxylin and Eosin and WSI: whole slide image
Pathologic Parameters
Tumor size, number of UMP, tumor sampling, questionable foci of CI or AI, nuclear atypia, oncocytic changes, solid growth pattern, mitotic activity, associated carcinomas, lymph node status, surgical margins and staging were all obtained from surgical pathology reports. Tumors were classified into three categories. The first category included tumors with questionable AI with or without questionable CI (n = 66), while the second category comprised tumors with questionable CI only (n = 258). UMP diagnosis was rendered only after examining additional deeper tissue levels in areas questionable for CI or AI (range: 2–10 deeper levels per suspicious focus, median 4).
Consistent with WHO classification, tumors with PTC-like nuclei were classified as WDT-UMP, whereas those with typical follicular nuclei were classified as follicular tumors of uncertain malignant potential (FT-UMP). Tumors with questionable CI or AI demonstrating oncocytic morphology were classified as OFT-UMP, a recognized subtype within FT-UMP. For the purposes of this study, OFT-UMP was analyzed distinctly.
The study cohort is divided into three analytic groups: FT-UMP, WDT-UMP and OFT-UMP. All tumor capsules were entirely submitted for histological examination and diagnosed or reviewed by senior endocrine pathologists (WCF and PMS) at the time of diagnosis. In our practice, “questionable” capsular invasion or angioinvasion refers to foci that raise suspicion for invasion but lack definitive histologic evidence after careful evaluation. In such cases, additional deeper levels (2–10 levels) were routinely examined to clarify the findings, rarely with additional ancillary testing. Diagnosis was ultimately based on histologic assessment by experienced endocrine pathologists, with consensus review performed in a subset of challenging cases (89 cases; 28%). Diagnostic criteria for each UMP category are summarized in Table 1.
Table 1.
Diagnostic criteria for uncertain malignant potential subtypes
| Classification | Questionable CI or AI | PTC-like nuclear atypia scorea | Oncocytic changes |
|---|---|---|---|
| WDT-UMP | Identified | 2–3 | Absent or focal |
| FT-UMP | Identified | 0–1 | Absent |
| OFT-UMP | Identified | 0–1 | Multifocal or diffuse |
PTC-nuclear atypia score 0–1 (no to minimal atypia), and score 2–3 (moderate to overt atypia) [17]
WDT-UMP, well-differentiated tumor of uncertain malignant potential; FT-UMP, follicular tumor of uncertain malignant potential; OFT-UMP, oncocytic (formerly Hürthle cell) follicular tumor of uncertain malignant potential; CI, capsular invasion; AI, angioinvasion; PTC,papillary thyroid carcinoma
Clinical and Molecular Parameters
Clinical data were obtained from the electronic medical record (Epic Systems, Verona, WI). The variables collected included patient age, sex, and tumor location, as well as preoperative FNAC results, which were categorized according to the Bethesda System for Reporting Thyroid Cytopathology [20]. Postoperative thyroid function studies, including thyroid stimulating hormone (TSH) and thyroglobulin levels (Tg), were recorded. The number of postoperative ultrasounds (US) was noted, and the extent of surgery was documented as total thyroidectomy (TT), hemithyroidectomy (HT), or completion thyroidectomy (CT), with HT defined as lobectomy with or without isthmusectomy. In addition, preoperative and postoperative molecular testing data (120 and 11 patients; respectively) were recorded, including results from Afirma Genomic Sequencing Classifier (GSC) and Xpression Atlas (XA) (Veracyte, South San Francisco, CA, USA), targeted next-generation sequencing with ThyroSeq (CBLPath, Rye Brook, NY, USA), miRNA gene expression and somatic gene alteration testing with ThyraMIR & ThyGeNEXT (Interpace Biosciences, Parsippany, NJ, USA), Genexus NGS, SNAPSHOT, PGDX2 and Solid Fusion Assay NGS (Thermo Fisher Scientific, Waltham, MA, USA). Information regarding radioactive iodine treatment follow-up and tumor recurrence or metastatic status were also collected for all patients. Follow-up with ≥ 6 months was considered.
Results
UMP accounted for the diagnosis in 0.5% of all thyroidectomies. The clinicopathologic and molecular characteristics are summarized in Table 2. Most patients (69%, n = 219) underwent HT, 29% (n = 93) underwent TT, and 2% (n = 6) had CT. The median age at diagnosis was 53-years (range 16–89) with a female-to-male ratio of 2:1. Among the UMP subtypes, OFT-UMP was most common (40%, n = 129), followed by FT-UMP (31%, n = 112) and WDT-UMP (22%, n = 83). The median tumor size was 3.2 cm (range 0.4–10 cm). Tumor size and Bethesda category for each UMP subtype are summarized in Table 3. Molecular aberrations were noted in 84% (117/140) of cases, with specific alterations identified in 62% (n = 72). RAS variants were most frequent (n = 56), followed by RAS-like alterations, gene fusions and PTEN loss by immunohistochemistry (3 cases each). The remaining cases were designated by Afirma as suspicious (n = 33) on Genomic Sequencing Classifier (GSC) without detailed mutational profile (resulted prior to Xpression Atlas), and ThyroSeq harboring a suspicious, unspecified gene expression profile (n = 2) or copy number alterations (n = 10). The detailed molecular testing for each UMP group is summarized in Table 4. Among cases with provided risk designations (n = 122), the risk of malignancy was low in 19% (n = 23), intermediate in 14% (n = 17), intermediate-high in 61% (n = 15), and high in 6% (n = 7). Clinical follow-up was available for 103 patients, with a median of 1.91 years (range 6 months-15.6 years) including 38 patients with more than 3 years of follow-up (median 5.3 years). No recurrences or metastases were observed. Of 103 patients with follow-up ≥ 6 months, 66% underwent postoperative ultrasound (n = 68).
Table 2.
Clinicopathologic and molecular characteristics of the Uncertain Malignant Potential patient cohort (318 patients/324 tumors)
| Clinicopathologic Characteristics | Variables | Total (N = 324) n/N (%) |
|---|---|---|
| Age, mean | Range (year) | 55 (16–89) |
| Sex | Female | 207/318 (65%) |
| Male | 111/318 (35%) | |
| Size of tumor, median | Range (cm) | 3.2 (0.4–10.0) |
| ≤ 1.0 | 15/321 (5%) | |
| 1.1–2.0 | 69/321 (21%) | |
| 2.1–4.0 | 131/321 (41%) | |
| > 4.0 | 106/321 (33%) | |
| Number of tumors | 1 | 312/318 (98%) |
| 2 | 6/318 (2%) | |
| FNAC | Yes | 202/324 (62%) |
| No | 122/324 (38%) | |
| Bethesda category of tumor | Nondiagnostic | 7/202 (3%) |
| Benign | 38/202 (19%) | |
| AUS | 102/202 (51%) | |
| FN | 42/202 (21%) | |
| Suspicious for malignancy | 10/202 (5%) | |
| Malignant | 3/202 (1%) | |
| Available molecular data | Available | 140/318 (44%) |
| Not available | 178/318 (56%) | |
| Positive molecular testing | Yes | 117/140 (84%) |
| No mutation (benign or negative) | 23/140 (16%) | |
| Reported genetic variant | Yes | 72/117 (62%) |
| No | 45/117 (38%) | |
| Gene reported | RAS genes (NRAS, HRAS, KRAS) | 57/72 (79%) |
| RAS-like genesa | 3/72 (4%) | |
| Gene fusionb | 3/72 (4%) | |
| PTEN c | 3/72 (4%) | |
| Other reported genetic variantsd | 6/72 (9%) | |
| pTERT p.C228T (3 cases) or TERT loss (1 case) | 4/140 (3%) | |
| No specific gene-level alteration reported | Copy number alteration by Thyroseq | 10/45 (22%) |
| Gene expression profile by Thyroseq | 2/45 (5%) | |
| Suspicious Afirmae | 33/45 (73%) | |
| Risk of malignancyf | 1–9% (low) | 23/122 (19%) |
| 10–49% (intermediate) | 17/122 (14%) | |
| 50–80% (intermediate-high) | 75/122 (61%) | |
| > 80% (high) | 7/122 (6%) | |
| Type of surgery | Total thyroidectomy | 93/318 (29%) |
| Hemithyroidectomy/Lobectomy | 219/318 (69%) | |
| Completion thyroidectomy | 6/318 (2%) | |
| Associated with carcinoma elsewhere in the thyroid gland | Yes | 88/318 (28%) |
| Postoperative initial US | 6 months to 1 year | 11/68 (16%) |
| 1–2 years | 24/68 (35%) | |
| After 2 years | 33/68 (49%) | |
| Follow-up US | Two or more postoperative US | 6/68 (9%) |
| Postoperative TSH levels | No suppression (TSH: >2 mlU/mL) | 25/110 (23%) |
| Low-normal (TSH: 0.5–2 mlU/mL) | 58/110 (53%) | |
| Mildly suppressed (TSH: 0.1–0.5 mlU/mL) | 14/110 (13%) | |
| Suppressed (TSH: <0.1 mlU/mL) | 13/110 (12%) | |
| Postoperative thyroglobulin levels | Undetectable (< 0.2–0.3 ng/ml) | 27/61 (44%) |
| Low-stable (0.3–1 ng/ml) | 8/61 (13%) | |
| Detectable (> 1 ng/ml) | 26/61 (43%) | |
| UMP classification | OFT-UMP | 129/324 (40%) |
| FT-UMP | 112/324 (34%) | |
| WDT-UMP | 83/324 (26%) |
RAS-like genes: 2 BRAF p.K601E and 1 EIF1X non-spice site variant
Gene fusions: 1 EML4::ALK and 2 PPARG (1 with PAX8 and 1 with unknown partner)
PTEN loss by immunohistochemistry
Other mutations include: MET overexpression, TSC2 p.P1458T, PIK3CA p.H1047Y, MAP2K1 p.Q243H & p.S243C, TP53 p.G262EFs*9, ETV6 p.A28T, JAK3 p.R395H, MLH3 p.K619I, DICER1 p.D1709E
Suspicious Afirma Genomic Sequencing Classifier (GSC) without detailed mutational profile
Risk of malignancy based on molecular testing (Thyroseq & Afirma) performed on fine needle aspiration cytology
FNAC, fine needle aspiration cytology; AUS, atypia of undetermined significance; FN, follicular neoplasm; US, ultrasound; TSH, thyroid stimulating hormone
Table 3.
Tumor size and Bethesda System Categories for Uncertain Malignant Potential subtypes
| Tumor Classification | Size, percentage of cases % (number of cases) | |||||
| ≤1.0 cm | 1.1–2.0 cm 2.1–4.0 cm | >4.0 cm | - | Total cases | ||
| WDT-UMP | 10% (8) | 22% (18) 30% (25) | 37% (31) | 1% (1) | 83 | |
| FT-UMP | 9% (10) | 15% (17) 39% (44) | 36% (40) | 1% (1) | 112 | |
| OFT-UMP | 8% (10) | 22% (28) 45% (59) | 24% (31) | 1% (1) | 129 | |
| Bethesda System Category, percentage of cases % (number of cases) | ||||||
| I | II III | IV | V | VI | Total cases | |
| WDT-UMP | 5% (3) | 13% (7) 56% (31) | 13% (7) | 9% (5) | 4% (2) | 55 |
| FT-UMP | 3% (2) | 27% (19) 54% (38) | 13% (9) | 3% (2) | 1% (1) | 71 |
| OFT-UMP | 3% (2) | 16% (12) 43% (33) | 34% (26)a | 4% (3) | - | 76 |
Among the 26 OFT-UMP classified as Bethesda IV, 15 cases were diagnosed as oncocytic neoplasm or suspicious for oncocytic neoplasm
WDT-UMP, well-differentiated tumor of uncertain malignant potential; FT-UMP, follicular tumor of uncertain malignant potential; OFT-UMP, oncocytic (formerly Hürthle cell) follicular tumor of uncertain malignant potential, “-”: missing value
Table 4.
Mutational landscape of the thyroid Tumors of Uncertain Malignant Potential cohort
| Gene mutation | Gene and variant | WDT-UMP | FT-UMP | OFT-UMP | Total | Note |
|---|---|---|---|---|---|---|
| RAS | NRAS p.Q61R | 6 | 8 | 6 | 20 | One OFT-UMP and one WDT-UMP plus TERT p.C288T |
| NRAS p.Q61K | 2 | 2 | 1 | 5 | One WDT-UMP plus EIF1AX p.A113_ splice variant | |
| NRAS p.G12A | - | 1 | - | 1 | plus NOTCH1 loss, RET loss, TERT loss | |
| NRAS a | 2 | 2 | 1 | 5 | ||
| HRAS p.Q61R | 4 | 2 | 1 | 7 | ||
| HRAS p.Q61K | 2 | 1 | - | 3 | One WDT-UMP plus EIF1AX p.A113 splice | |
| HRAS p.G13R | - | - | 1 | 1 | ||
| KRAS p.Q61R | 1 | 1 | - | 2 | ||
| KRAS p.G12V | 1 | - | 3 | 4 | One OFT-UMP plus TERT p.C288T | |
| KRAS a | 1 | - | - | 1 | ||
| RAS p.Q61Ra | 4 | 1 | 3 | 8 | ||
| RAS-like | BRAF p.K601E | 1 | 1 | - | 2 | |
| EIF1AX non-splice site variant | - | - | 1 | 1 | ||
| Gene fusion | EML4::ALK | 1 | - | - | 1 | |
| PAX8::PPARG | - | - | 1 | 1 | ||
| PPARG b | - | 1 | - | 1 | ||
| PTEN | PTEN lossc | - | 1 | 2 | 3 | |
| Other mutations | MET overexpression | - | - | 1 | 1 | |
| TSC2 p.P1458T | - | - | 1 | 1 | ||
| PIK3CA p.H1047Y | 1 | - | - | 1 | plus DICER1 p.D1709E | |
| MAP2K1 p.Q243H & p.S243C | 1 | - | - | 1 | ||
| TP53 p.G262EFs*9 | - | - | 1 | 1 | plus copy number alteration | |
| ETV6 p.A28T | - | 1 | - | 1 | plus JAK3 p.R395H & MLH3 p.K619I | |
| Abnormal with no | Copy number alteration | 2 | 4 | 4 | 10 | |
| reported specific | Gene expression profile | 2 | - | - | 2 | |
| mutation | Suspicious Afirmad | 9 | 9 | 15 | 33 | |
| Benign | Benign Afirma/Thyroseq | 3 | 9 | 11 | 23 | |
| Total | 43 | 44 | 53 | 140 |
RAS variant is not specified but RAS p.Q61R is positivity is detected by immunohistochemistry (does not distinguish NRAS/HRAS/KRAS)
PPARG overexpression potentially indicative of fusion
PTEN loss by immunohistochemistry
Suspicious Afirma Genomic Sequencing Classifier (GSC) without detailed mutational profile
WDT-UMP, well-differentiated tumor of uncertain malignant potential; FT-UMP, follicular tumor of uncertain malignant potential; OFT-UMP, oncocytic (formerly Hürthle cell) follicular tumor of uncertain malignant potential, “-”: missing value
Discussion
In the WHO classification of tumors, NIFTP, FT- UMP, and WDT-UMP are categorized as low-risk, follicular-derived neoplasms, reflecting their low likelihood for metastasis or structural recurrence [19]. Although familiar with the implications for a NIFTP diagnosis, most pathologists in practice, particularly in the United States, are hesitant to diagnose UMP, despite its routine use in other organ systems (e.g. Uterine Smooth Muscle Tumor: STUMP; Melanocytic Tumor: MELTUMP). Inherent in the UMP name, the diagnosis is equivocal for malignancy, and such a diagnosis of uncertainty is anxiety-provoking for patients and physicians. However, biology is complex, and with each scientific breakthrough in our fundamental understanding of a biological process, there is concomitant nuance that appends that diagnostic certainty with the words “in most cases.” Furthermore, recent reports indicate low reproducibility in diagnosing CI and AI in encapsulated follicular-patterned tumors, particularly in cases that are diagnostically challenging [21, 22], and the term UMP may be usefully applied to at least some of them.
The UMP designation is applied to tumors with questionable foci of CI and/or AI despite extensive evaluation, including multiple hematoxylin and eosin-stained deeper levels on individual tissue blocks [19]. UMP terminology is rarely utilized, with reported incidence ranging from 0.5% to 3% in most published series, consistent with our cohort’s reported incidence of 0.5% [23–26]. The age range for UMP cases was broad (16–89 years), with a median age of 53 years, consistent with prior reports of 50–55 years [23, 26]. UMP is reported approximately twice as often in women [23, 26]. The mean size of UMP is 3.2 cm, which is consistent with that reported by the cohort of Cracolici et al. (3.4 cm) [10], with the majority of tumors falling in the 2.1–4.0 cm range across all subtypes and most UMP were > 2 cm (72%; n = 259) (Table 3). Prior reports show no significant differences between nodule size or volume change and the final diagnoses of benign, UMP, or malignant [27]. The moderate size of UMP nodules may nonetheless prompt surgical excision or possible total thyroidectomy. This highlights that nodule size alone, particularly circumscribed nodules, should not be overinterpreted as a predictor of malignancy nor used as a sole indication for total thyroidectomy.
Cytologic diagnosis is key in evaluating thyroid nodules. In our UMP cohort, WDT-UMP and FT-UMP were most often classified as Bethesda III (56% and 54%, respectively), and OFT-UMP as Bethesda IV (34%), highlighting the predominance of intermediate Bethesda categories across the spectrum (Table 3). These findings align with a prior study of 28 UMP cases, in which most cases also fell into Bethesda III–IV categories, supporting the consistency of cytologic patterns despite smaller sample sizes [28]. Although cytologic diagnosis guides triage for molecular testing and surgical decisions, it stumbles with follicular thyroid neoplasia, requiring histological assessment following resection [29]. This series represents the largest collection of UMP with comprehensive molecular analysis, performed in 140 of 318 patients. RAS and RAS-like alterations promote follicular thyroid neoplasia, including UMP. Duan et al. studied 17 WDT-UMP and 32 FT-UMP and found that RAS (NRAS/HRAS/KRAS) mutations are seen in 53% of WDT-UMP and 27% of FT-UMP [30] which is consistent with our results for WDT-UMP (53%; 23/43), FT-UMP (41%; 18/44) and OFT-UMP (28%; 15/53). EIF1AX mutations have been reported in FT-UMP, and one of our WDT-UMP cases demonstrated concomitant RAS and EIF1AX mutations. These alterations are known to occur in follicular-patterned thyroid neoplasms and have been documented to co-exist within the same tumor [30, 31]. Although not previously described in UMP, BRAF p.K601E and PPARG fusions are regarded as RAS-like alterations [31, 32] and were seen in 3 UMP. EML4::ALK fusion has been reported in kinase fusion-related thyroid carcinomas and has been recognized as a driver of some aggressive thyroid tumors [33], with EML4::ALK-positive thyroid carcinomas demonstrating clinical response to ALK inhibitors [33]. In our cohort, one UMP harboring EML4::ALK (3 cm) showed PTC-like nuclear atypia with questionable foci of CI and was classified as WDT-UMP.
Molecular changes traditionally associated with carcinoma are particularly problematic for pathologists to classify if they have the molecular data in hand. Pathologists may consider classifying tumors with EML4::ALK fusions as encapsulated follicular variant of PTC (EFVPTC) or encapsulated solid PTC when the solid component exceeds 50% [19]. Three UMP cases harbored TERT promoter variants with concurrent RAS driver mutations, consistent with prior reports of TERT promoter alterations in FT-UMP [34, 35]. We note that TERT promoter mutations may be subclonal, and because our molecular analyses were performed on either FNA material or a single block without specific TERT targeting, such mutations could have been missed and TERT promoter variants may occur with greater frequency in UMP [36]. In one study, two FT-UMP cases with TERT promoter alterations (one with concomitant NRAS, 3.5 cm; and one isolated TERT promoter, 4.8 cm), both with the entire capsule submitted, showed no recurrence or distant metastasis, although follow-up was less than two years for both tumors [34]. In the series by Hysek et al., TERT promoter alterations were identified in 16% of FT-UMP (8/52), with three patients developing distant metastases (two within three years and one at six years), requiring additional radioactive iodine therapy. Notably, tumor size was reported in only two of these cases (7 cm and 4 cm), and extent of capsule/tumor interface submission was unreported [35]. Importantly, while TERT promoter mutations have been associated with aggressive thyroid carcinomas, in our opinion, their presence alone should not be used to diagnose carcinoma in the absence of histologic evidence of invasion. PTEN loss has not been reported in UMP despite its variable association with thyroid neoplasia [37]. In our cohort, three UMP patients with PTEN loss by immunohistochemistry presented with multiple nodules. One patient had confirmed PTEN hamartoma tumor syndrome (PHTS) and PTEN aberration, while the other two patients, considered clinically sporadic tumors, did not undergo molecular testing to confirm PTEN alteration. TP53 mutations frequently coexist with other driver mutations in thyroid neoplasia and are associated with aggressive tumor behavior [38]. Consistent with our findings, TP53 alterations may also be present in borderline tumors, and one study reported TP53 mutations in 2 cases of FT-UMP, and an isolated TP53 mutation was observed in one WDT-UMP [30]. Preoperative ThyroSeq detected a suspicious but unspecified gene expression profile in 2 UMP and copy number alterations in 10 UMP. Similar alterations have been described in EFVPTC and minimally invasive follicular thyroid carcinomas [39]. Other rare alterations identified in UMP are summarized in Table 4. These include MET overexpression, TSC2 variants or MAP2K1 mutations, which have not previously been reported in UMP; however, they have been previously described in papillary thyroid carcinomas, including more aggressive cases [40–42]. A DICER1 mutation (codon 1709) was also identified in one UMP. DICER1 is a recognized driver in thyroid neoplasias and tumor-like conditions, including pediatric high-grade thyroid carcinoma, adolescent follicular nodular disease, and macrofollicular subtypes of follicular carcinoma [43, 44]. In our UMP cohort, it occurred as a rare finding and was identified in conjunction with a PIK3CA p.H1047Y mutation. Given the small sample size and limited follow-up, no definitive conclusions can be drawn about the potential behavior of UMP harboring these mutations. The detection of such mutations in a follicular-patterned tumor should prompt careful examination of the entire lesional periphery to exclude invasion.
The presence of putative high-risk molecular alterations in encapsulated, noninvasive tumors is disconcerting and may suggest that these nodules acquire driver events before developing overt architecturally invasive features. This may promote more aggressive clinical management despite a diagnosis of UMP [34, 35, 39, 45].
Although postoperative TSH levels, serum thyroglobulin (Tg) levels, and ultrasound follow-up data were collected for this cohort, the interpretability of these parameters is limited, as most patients underwent hemithyroidectomy. Consequently, while the availability of these data is novel, definitive conclusions regarding postoperative biochemical or radiologic outcomes cannot be drawn.
UMP is a rarely utilized but clinically important diagnostic category that shows overwhelmingly indolent behavior, even in tumors with high-risk molecular alterations. In this large UMP cohort with clinical follow-up extending up to 15 years, we observed no recurrences or metastases, supporting conservative management (hemithyroidectomy), in line with current American Thyroid Association guidelines [46]. Total thyroidectomy incurs variable risks of nerve injury, hypoparathyroidism, and assured need for thyroid hormone replacement. We note that in this cohort, total thyroidectomy was performed in 29% of patients, at least a subset of which were likely overtreated in the context of their biological risk for recurrence.
This study is limited by the relative diagnostic novelty of UMP terminology and its employment by pathologists, incomplete long-term follow-up for patients diagnosed primarily with UMP, and variable availability of established prognostic markers like Ki-67 and molecular testing across the long duration of this study period. Although the WHO 5th edition appropriately categorizes UMP within the spectrum of low-risk neoplasms, the inclusion of the word “malignant” as part of the diagnostic designation poses challenges for patient counseling and shared decision-making. As illustrated by NIFTP, meaningful refinement of thyroid tumor classification/nomenclature depends on continued correlation of histologic findings with molecular alterations and long-term clinical outcomes. Here, evolution from UMP toward tumors of low biological potential (LBP) may eventually be considered. As noted by the conglomerate of morphologic and molecular changes that may be stratified to UMP, individual cohorts depicting isolated morphologic or molecular changes will need to be studied in finer detail. Nonetheless, the consistency of findings across histologic subtypes and molecular profiles strengthens the conclusion that UMP represents a biologically low-risk group of tumors. Larger, multi-institutional studies with longer-term follow-up are needed to further refine risk stratification and clinical management.
Acknowledgements
Data from this manuscript was presented, in part, at the 115th annual meeting of the United States and Canadian Academy of Pathology on March 24, 2026, in San Antonio, TX, USA.
Funding
TSA was supported by grant number T32GM144273 from the National Institute of General Medical Sciences.
Footnotes
Competing Interests Dr. Sadow is a member of the Endocrine Pathology editorial board as disclosed in the manuscript.This is the only competing interest of any of the authors.
Disclaimer The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Institutes of Health.
Author Disclosure Statement Dr. Peter Sadow is a member of the Endocrine Pathology editorial board. This article was handled by an independent senior editor and was peer-reviewed according to the journal’s standard procedures. The remaining authors declare no competing interests.
Data Availability
Data has been generated from the primary work of authors and is presented in the manuscript per Mass General Brigham Internal Review Board Guidelines (to PMS, 2011P000013). Any related data can be provided as requested.
References
- 1.Graham A Malignant Tumors of the Thyroid: Epithelial Types. Ann Surg. 1925;82(1):30–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lindsay S Carcinoma of the thyroid gland: a clinical and pathologic study of 293 patients at the University of California hospital Springfield, Illinois, USA: Charles C. Thomas; 1960. [Google Scholar]
- 3.Lang W, Georgii A, Stauch G, Kienzle E. The differentiation of atypical adenomas and encapsulated follicular carcinomas in the thyroid gland. Virchows Arch A Pathol Anat Histol. 1980;385(2):125–41. [DOI] [PubMed] [Google Scholar]
- 4.Hazard JB, Kenyon R. Encapsulated angioinvasive carcinoma (angioinvasive adenoma) of thyroid gland. Am J Clin Pathol. 1954;24(7):755–66. [DOI] [PubMed] [Google Scholar]
- 5.Hazard JB, Kenyon R. Atypical adenoma of the thyroid. AMA Arch Pathol. 1954;58(6):554–63. [PubMed] [Google Scholar]
- 6.Meissner WA, Warren S. Tumors of the thyroid gland. In: Atlas of tumor pathology, Second Series, Fascicle 4, (Eds.). Washington, DC, USA: Armed Forces Institute of Pathology (AFIP), 1969:1–135. [Google Scholar]
- 7.Woolner LB. Thyroid carcinoma: pathologic classification with data on prognosis. Semin Nucl Med. 1971;1(4):481–502. [DOI] [PubMed] [Google Scholar]
- 8.Franssila KO, Ackerman LV, Brown CL, Hedinger CE. Follicular carcinoma. Semin Diagn Pathol. 1985;2(2):101–22. [PubMed] [Google Scholar]
- 9.Williams ED. Guest Editorial: Two Proposals Regarding the Terminology of Thyroid Tumors. Int J Surg Pathol. 2000;8(3):181–3. [DOI] [PubMed] [Google Scholar]
- 10.Cracolici V, Ritterhouse LL, Segal JP, Puranik R, Wanjari P, Kadri S, 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):881–92. [DOI] [PubMed] [Google Scholar]
- 11.Tzen CY, Huang YW, Fu YS. Is atypical follicular adenoma of the thyroid a preinvasive malignancy? Hum Pathol. 2003;34(7):666–9. [DOI] [PubMed] [Google Scholar]
- 12.Okayasu I, Osakabe T, Fujiwara M, Fukuda H, Kato M, Oshimura M. Significant correlation of telomerase activity in thyroid papillary carcinomas with cell differentiation, proliferation and extra-thyroidal extension. Jpn J Cancer Res. 1997;88(10):965–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Giardina C, Pollice L, Ricco R, Vacca E, Pennella A, Serio G, et al. Differential diagnosis between thyroid follicular adenoma and carcinoma. Analytic morphometric approach. Pathol Res Pract. 1989;185(5):726–8. [DOI] [PubMed] [Google Scholar]
- 14.Rivera M, Tuttle RM, Patel S, Shaha A, Shah JP, Ghossein RA. Encapsulated papillary thyroid carcinoma: a clinico-pathologic study of 106 cases with emphasis on its morphologic subtypes (histologic growth pattern). Thyroid. 2009;19(2):119–27. [DOI] [PubMed] [Google Scholar]
- 15.Nikiforova MN, Nikiforov YE. Molecular genetics of thyroid cancer: implications for diagnosis, treatment and prognosis. Expert Rev Mol Diagn. 2008;8(1):83–95. [DOI] [PubMed] [Google Scholar]
- 16.Hirokawa M, Carney JA, Goellner JR, DeLellis RA, Heffess CS, Katoh R, et al. Observer variation of encapsulated follicular lesions of the thyroid gland. Am J Surg Pathol. 2002;26(11):1508–14. [DOI] [PubMed] [Google Scholar]
- 17.Nikiforov YE, Seethala RR, Tallini G, Baloch ZW, Basolo F, Thompson LD, et al. Nomenclature Revision for Encapsulated Follicular Variant of Papillary Thyroid Carcinoma: A Paradigm Shift to Reduce Overtreatment of Indolent Tumors. JAMA Oncol. 2016;2(8):1023–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lloyd RV, Osamura RY, Klöppel G, Rosai J. WHO Classification of Tumours of Endocrine Origin. 4th ed. Lyon, France: International Agency for Research on Cancer (IARC); 2017. [Google Scholar]
- 19.Baloch ZW, Asa SL, Barletta JA, Ghossein RA, Juhlin CC, Jung CK, et al. Overview of the 2022 WHO Classification of Thyroid Neoplasms. Endocr Pathol. 2022;33(1):27–63. [DOI] [PubMed] [Google Scholar]
- 20.Cibas ES, Ali SZ. The 2017 Bethesda System for Reporting Thyroid Cytopathology. Thyroid. 2017;27(11):1341–6. [DOI] [PubMed] [Google Scholar]
- 21.Mansour Y, Chaltiel L, Cavillon A, Al Bouzidi A, Kamate B, De Leo A, et al. Can we improve the diagnosis of invasion in encapsulated follicular-patterned thyroid tumors? Data from a massive international e-learning initiative. Virchows Arch. 2025;487(1):105–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Zhu Y, Li Y, Jung CK, Song DE, Hang JF, Liu Z, et al. Histopathologic Assessment of Capsular Invasion in Follicular Thyroid Neoplasms-an Observer Variation Study. Endocr Pathol. 2020;31(2):132–40. [DOI] [PubMed] [Google Scholar]
- 23.Baser H, Topaloglu O, Tam AA, Alkan A, Kilicarslan A, Ersoy R, et al. Comparing Clinicopathologic and Radiographic Findings Between TT-UMP, Classical, and Non-Encapsulated Follicular Variants of Papillary Thyroid Carcinomas. Endocr Pathol. 2016;27(3):233–42. [DOI] [PubMed] [Google Scholar]
- 24.Liu Z, Zhou G, Nakamura M, Koike E, Li Y, Ozaki T, et al. Encapsulated follicular thyroid tumor with equivocal nuclear changes, so-called well-differentiated tumor of uncertain malignant potential: a morphological, immunohistochemical, and molecular appraisal. Cancer Sci. 2011;102(1):288–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Piana S, Frasoldati A, Di Felice E, Gardini G, Tallini G, Rosai J. Encapsulated well-differentiated follicular-patterned thyroid carcinomas do not play a significant role in the fatality rates from thyroid carcinoma. Am J Surg Pathol. 2010;34(6):868–72. [DOI] [PubMed] [Google Scholar]
- 26.Hofman V, Lassalle S, Bonnetaud C, Butori C, Loubatier C, Ilie M, et al. Thyroid tumours of uncertain malignant potential: frequency and diagnostic reproducibility. Virchows Arch. 2009;455(1):21–33. [DOI] [PubMed] [Google Scholar]
- 27.Nakamura H, Hirokawa M, Ota H, Kihara M, Miya A, Miyauchi A. Is an Increase in Thyroid Nodule Volume a Risk Factor for Malignancy? Thyroid. 2015;25(7):804–11. [DOI] [PubMed] [Google Scholar]
- 28.Zylka A, Dobruch-Sobczak K, Piotrzkowska-Wroblewska H, Jedrzejczyk M, Goralski P, Galczynski J, et al. Ultrasound and cytopathological characteristics of thyroid tumours of uncertain malignant potential - from diagnosis to treatment. Endokrynol Pol. 2024;75(2):170–8. [DOI] [PubMed] [Google Scholar]
- 29.Kakudo K How to handle borderline/precursor thyroid tumors in management of patients with thyroid nodules. Gland Surg. 2018;7(Suppl 1):S8–S18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Duan H, Liu X, Ren X, Zhang H, Wu H, Liang Z. Mutation profiles of follicular thyroid tumors by targeted sequencing. Diagn Pathol. 2019;14(1):39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Cancer Genome Atlas Research N. Integrated genomic characterization of papillary thyroid carcinoma. Cell. 2014;159(3):676–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.De Leo A, Ruscelli M, Maloberti T, Coluccelli S, Repaci A, de Biase D, et al. Molecular pathology of endocrine gland tumors: genetic alterations and clinicopathologic relevance. Virchows Arch. 2024;484(2):289–319. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Demeure MJ, Aziz M, Rosenberg R, Gurley SD, Bussey KJ, Carpten JD. Whole-genome sequencing of an aggressive BRAF wild-type papillary thyroid cancer identified EML4-ALK translocation as a therapeutic target. World J Surg. 2014;38(6):1296–305. [DOI] [PubMed] [Google Scholar]
- 34.Alzumaili BA, Instrum R, Alabkaa A, Sadow PM, Tuttle MR, Xu B, et al. Pathological Diagnosis of Thyroid Nodules with Preoperatively Detected TERT Promoter Mutations in the Absence of BRAF(V600E): A Bi-Center Series of 52 Cases. Thyroid. 2025;35(10):1145–52. [DOI] [PubMed] [Google Scholar]
- 35.Hysek M, Paulsson JO, Jatta K, Shabo I, Stenman A, Hoog A, et al. Clinical Routine TERT Promoter Mutational Screening of Follicular Thyroid Tumors of Uncertain Malignant Potential (FT-UMPs): A Useful Predictor of Metastatic Disease. Cancers (Basel). 2019;11(10):1443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Hysek M, Jatta K, Hellgren LS, Stenman A, Larsson C, Zedenius J, et al. Spatial Distribution Patterns of Clinically Relevant TERT Promoter Mutations in Follicular Thyroid Tumors of Uncertain Malignant Potential: Advantages of the Digital Droplet PCR Technique. J Mol Diagn. 2021;23(2):212–22. [DOI] [PubMed] [Google Scholar]
- 37.Plitt G, Brewer T, Yehia L, Rabinowitz L, Griffith CC, Eng C. The Genomic Landscape of Benign and Malignant Thyroid Tumors from Individuals Carrying Germline PTEN Variants Is Distinct from Sporadic Thyroid Cancers. Cancer Res. 2024;84(21):3657–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Ibrahimpasic T, Ghossein R, Shah JP, Ganly I. Poorly Differentiated Carcinoma of the Thyroid Gland: Current Status and Future Prospects. Thyroid. 2019;29(3):311–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Brandler TC, Zhou F, Liu CZ, Serrano A, Sun W, Nikiforov YE, et al. Molecular Profiles of Noninvasive, Minimally Invasive, and Invasive Follicular Patterned Thyroid Neoplasms with Papillary Nuclear Features. Thyroid. 2023;33(6):715–23. [DOI] [PubMed] [Google Scholar]
- 40.Xu R, Wan Y, Ding B. Overexpressed MET drives aggressive thyroid cancer phenotypes and serves as a precision therapeutic target. Sci Rep. 2025;15(1):39809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Derwich A, Sykutera M, Brominska B, Andrusiewicz M, Ruchala M, Sawicka-Gutaj N. Clinical Implications of mTOR Expression in Papillary Thyroid Cancer-A Systematic Review. Cancers (Basel). 2023;15(6):1665. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Bu R, Siraj AK, Masoodi T, Parvathareddy SK, Iqbal K, Al-Rasheed M, et al. Recurrent Somatic MAP2K1 Mutations in Papillary Thyroid Cancer and Colorectal Cancer. Front Oncol. 2021;11:670423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Lee YA, Im SW, Jung KC, Chung EJ, Shin CH, Kim JI, et al. Predominant DICER1 Pathogenic Variants in Pediatric Follicular Thyroid Carcinomas. Thyroid. 2020;30(8):1120–31. [DOI] [PubMed] [Google Scholar]
- 44.Wasserman JD, Sabbaghian N, Fahiminiya S, Chami R, Mete O, Acker M, et al. DICER1 Mutations Are Frequent in Adolescent-Onset Papillary Thyroid Carcinoma. J Clin Endocrinol Metab. 2018;103(5):2009–15. [DOI] [PubMed] [Google Scholar]
- 45.Kim TH, Kim YE, Ahn S, Kim JY, Ki CS, Oh YL, et al. TERT promoter mutations and long-term survival in patients with thyroid cancer. Endocr Relat Cancer. 2016;23(10):813–23. [DOI] [PubMed] [Google Scholar]
- 46.Ringel MD, Sosa JA, Baloch Z, Bischoff L, Bloom G, Brent GA, et al. 2025 American Thyroid Association Management Guidelines for Adult Patients with Differentiated Thyroid Cancer. Thyroid. 2025;35(8):841–985. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data has been generated from the primary work of authors and is presented in the manuscript per Mass General Brigham Internal Review Board Guidelines (to PMS, 2011P000013). Any related data can be provided as requested.