What lymph node levels does a lateral [comprehensive / therapeutic] neck dissection for differentiated thyroid cancer (DTC) include?
Although the rate of clinical nodal involvement in the lateral compartment was initially described by Japanese (Noguchi et al. 1970) and Germans (Gimm et al. 1998):
Sivanandan et al (2001) were the first to systematize it by levels
In 2013, the Canadian group of Jeremy L. Freeman (Eskander et al. Thyroid) conducted a systematic review that included the meta-analysis of 18 publications (including his 2012 retrospective work with 185 patients; Merdad et al. Head Neck) agglutinating 1298 lateral neck dissections for DTC:
Although heterogeneity was a constant in all comparisons by levels (I2: 31% to 87%), it is the best evidence to date that justifies the use of selective emptying IIa-Vb in this cohort of patients with this pathology:
Level III is the most frequently compromised
The majority (73%) of patients have more than one level involved:
Level III and IV:
46%
Level II, III and IV:
26%
Level III, IV and V:
11%
Level II, III, IV and V:
13% (Merdad et al. 2012)
Levels I and sublevel Va (cranial to the distal spinal nerve pathway):
Are rarely involved, usually in patients with high disease volume and multilevel invasion
Emptying of sublevel IIb (retrospinal recess):
Is usually indicated when clinical, radiological or macroscopic involvement is evident intraoperatively
The macroscopic involvement evident in the intraoperative sublevel IIa usually determines the addition of sublevel IIb to the neck dissection
“Skip metastases” within the lateral compartment are uncommon and occur in around 9% of patients:
Level II with level III and IV
Level V with level III and IV
(Merdad et al. 2012)
Selective lymphadenectomy IIa to Vb currently dissects levels IIa, III, IV, Vb and the “infraspinal” portion of the VA [VAi] in order to avoid the functional sequelae of cranial nerve XI dissection
Key Sonographic Features Suggestive of Malignancy:
Based on contemporary radiology references:
Hypoechoic echotexture
Posterior acoustic shadowing
Irregular or spiculated margins
Angular or microlobulated margins
Taller-than-wide orientation:
Non-parallel to skin
Thick echogenic halo:
Suggests desmoplastic reaction
Ductal extension or branching pattern
Intra-lesional calcifications
Increased stiffness on elastography:
High shear-wave values
Low apparent diffusion doefficient (ADC) values on diffusion-weighted imaging (DWI)
These align with standard teaching but are now reinforced by quantitative imaging advances such as elastography and DWI
Caveats — Not Exclusively Malignant:
Well-defined smooth borders and posterior acoustic enhancement can appear in both benign and malignant lesions:
Interpretation must rely on the whole feature set
Layering or “teacup” micro-calcifications on mammography typically lean benign, despite ultrasound appearance; correlation remains essential
Integrated Imaging Approach:
A sonographic mass that appears benign on mammography should be evaluated primarily based on B-mode ultrasound features rather than mammographic impression
Use of color Doppler improves specificity for malignancy in non-mass-like lesions without reducing sensitivity
Quantitative Imaging Enhancements:
Diffusion-weighted imaging (DWI):
Apparent Diffusion Coefficient (ADC) values provide quantitative assessment:
Malignant lesions typically show mean ADC ≈ 1.03 ×10⁻³ mm²/s, benign ≈1.5 ×10⁻³ mm²/s; ADC <1.0 ×10⁻³ mm²/s strongly favors malignancy
Recent meta-analysis confirms ADC’s usefulness for distinguishing lesions, though exact thresholds vary:
Most protocols now use ≥ 1.5T MRI with b-values around 800 s/mm²
Shear-wave elastography (SWE):
Mean values for malignancies often exceed 133 to 153 kPa (e.g., ~167 kPa), aiding differentiation
Full References:
Malherbe K. Breast Ultrasound. StatPearls, updated 2024 – Highlights classic sonographic features: hypoechoic texture, shadowing, margins, etc Verywell Health, “Breast Cancer Ultrasound: How It Works and What Results Mean.” (2022) – Discusses overlap of benign/malignant features and interpretive context.
Tarigan VN et al. 2025 systematic review: DWI (ADC) helps distinguish benign vs malignant lesions; ADC measurement challenges remain. Kwon M et al. (2024) – Mean ADC ~0.982 ×10⁻³ mm²/s for cancers; SWE stiffness ~167.7 kPa. Surov A et al. (2019) – Pooled ADC values show malignant lesions average 1.03 ×10⁻³ mm²/s vs benign 1.5 ×10⁻³ mm²/s; benign rarely under 1.0 ×10⁻³ mm²/s.
Stavros AT. Breast Ultrasound, 2004 – Covers foundational ultrasound interpretation concepts. Cardenosa G. Clinical Breast Imaging: The Essentials, 2015 – Classic reference listing the ten ultrasound signs of malignancy.
Tarigan VN et al. (2024) Frontiers in Oncology – Color Doppler specificity enhancement in non-mass lesions.
Summary of the major updates in the WHO Classification of Tumors of the Breast, 6th Edition (2026) as they pertain to invasive lobular carcinoma (ILC) and related changes
Key ILC-Specific Changes:
ILC with Extracellular Mucin (ILCEM) — New Diagnostic Entity:
The most significant ILC-related change is the formal recognition of ILC with extracellular mucin (ILCEM) as a distinct diagnostic entity with prognostic implications:
This rare subtype, first described in 2009, is characterized by lobular-phenotype tumor cells (single cells, cords, nests, trabeculae) floating within pools of extracellular mucin
Key features include:
Typically presents as a large mass (> 2 cm) in postmenopausal women:
Often shows high nuclear grade (grade 2 to 3), signet-ring cell morphology, pleomorphic features, solid growth patterns
E-cadherin is absent or aberrant on IHC, with cytoplasmic p120 catenin localization confirming lobular phenotype in both mucinous and non-mucinous components
All reported cases are ER-positive, with a higher rate of HER2 positivity than classic ILC
Molecularly, ILCEM harbors CDH1 alterations in ~ 92% of cases, along with PIK3CA, RUNX1, AKT1, and PTEN mutations:
Cases with recurrences show additional ERBB2, ERBB3, TP53, and FGFR1 alterations
Worse prognosis than classic ILC:
52% recurrence rate and ~ 30% disease-specific mortality in reported series
The WHO 6th Edition (2026) introduces a critical terminological change for ILC:
The term “variant” is now reserved exclusively for molecular / genetic alterations:
While morphological differences in ILC are reclassified as architectural patterns or cytomorphological subtypes rather than “variants”;
This represents a significant conceptual shift from the 5th edition framework
Terminology Update: “Variant” vs. “Pattern/Subtype”:
In the WHO 5th edition, ILC morphological forms were called “variants”
The 6th edition clarifies that “variant” should now refer only to molecular / genetic alterations (e.g., CDH1 mutations, PIK3CA mutations):
While the morphological diversity of ILC is described using terms like “pattern” or “subtype”:
This aligns breast pathology terminology with broader WHO classification principles across organ systems
Architectural Patterns of ILC:
ILC architectural patterns are categorized based on growth pattern and cytomorphology:
Growth pattern-based:
Classic:
The most common pattern, characterized by small, dyscohesive cells with monomorphic nuclei and scant cytoplasm
Arranged in single cells, single files, and targetoid infiltrations around ducts and lobules, with little or no stromal reaction
Two cell types exist within classic ILC:
Type A cells – monomorphic, pleomorphism score 1
Type B cells – larger, vesicular, pleomorphism score 2
Solid:
Sheets of dyscohesive lobular cells growing in solid nests without fibrovascular cores or a fibrous capsule
Associated with higher Ki67 and more aggressive behavior
High-grade solid ILC is grouped with pleomorphic ILC as an aggressive subtype
Alveolar:
Cells arranged in rounded nests of ~ 20 cells resembling alveolar structures; often mixed with other patterns
Tubulolobular:
Features small tubular structures admixed with classic lobular single-file growth
Trabecular:
Trabeculae mainly 2 to 3 cells thick (first described by Martinez and Azzopardi in 1979)
Prognosis appears similar to classic ILC
Solid papillary:
A recently described pattern with circumscribed nodules containing fibrovascular cores, distinct from solid ILC (which lacks fibrovascular cores)
May express neuroendocrine markers (synaptophysin, chromogranin)
Shows higher post-endocrine therapy Ki67 levels
ILC with tubular elements (ILC-TE):
A recently identified pattern defined by noncohesive carcinoma cells mixed with cohesive tubular elements, with complete E-cadherin loss but P-cadherin upregulation (E-cadherin to P-cadherin switch, or EPS) in tubular areas
Accounts for ~ 7.5% of ILC and is associated with less-aggressive features (lower grade, lower Ki67, cT1, cN0)
Cytomorphology-based:
Pleomorphic:
High nuclear grade (pleomorphism score 3), accounting for < 1% of all breast cancers
Associated with worse prognosis than classic ILC and even IBC-NST, with frequent ERBB2 and PIK3CA mutations
Histiocytoid / apocrine:
Cells with abundant eosinophilic or granular cytoplasm resembling histiocytes
New Entity: ILC with Extracellular Mucin (ILCEM):
The 6th edition formally recognizes ILCEM as a distinct diagnostic entity:
Lobular-phenotype cells floating within pools of extracellular mucin, confirmed by absent /aberrant E-cadherin and cytoplasmic p120 catenin
This carries worse prognosis than classic ILC (52% recurrence rate)
Prognostic Stratification by Pattern:
A large cohort study (n = 7,140) identified that pleomorphic ILC and high-grade solid ILC together comprise ~ 14% of ILC cases and constitute an aggressive subtype with worse breast cancer-specific survival and disease-free survival compared to both classic ILC and IBC-NST:
Notably, adjuvant chemotherapy did not improve outcomes in this aggressive subgroup
Classic ILC and its related patterns (alveolar, trabecular, papillary, tubulolobular) had significantly better survival than IBC-NST in the first 10 to 15 years of follow-up
Significant increase in incidence after mammography screening:
10-fold to 17-fold increase from 1970’s to 2004:
1 in 1300 mammograms
Represents 20% to 25% of all screen-detected breast neoplasias diagnosed annually
Based on autopsy data:
The prevalence of undetected DCIS may exceed 20% in women over 40:
Underscoring the concern for overdiagnosis:
Modeling studies estimate up to 65% of DCIS may be overdiagnosed
DCIS by it self:
Is not a risk to life
DCIS may progress to invasion and compromise survival:
If left untreated:
1 in 6 DCIS patients:
Progress to invasive breast cancer (IBC):
70% estimated to remain indolent
At this time we do not have any robust biomarkers:
That can quantify the risk of progression to IBC or
Help us separate indolent disease:
From the potentially dangerous lesions
Risk of overtreament:
The increase incidence of DCIS in mammographically detected cases:
Has not lead to a decrease in the incidence of IBC or reduction of IBC morality
Risk factors for progression / recurrence of DCIS:
The risk factors for IBC recurrence may be different from the risks factors for DCIS recurrence?
Risk of IBC recurrence:
African American race
Premenopausal status
Detection by palpation
Involved margins
High histologic grade
High p16 expression
Risk of IBC or DCIS recurrence:
DCIS size
Histology type
Comedo necrosis
Grade (High)
Young age
Close margins or positive margins
Patients with DCIS that recur with an IBC:
Some patients with DCIS may develop:
Progression of there disease:
30% (1 out of 6)
Some will have a de novo invasive breast cancer
Some might have a missed invasive cancer?
Patient that have a DCIS recurrence:
Might be a true in situ recurrence
De novo DCIS
Residual disease?
Studies are describing observations of events like:
Synchronous IBC
Subsequent ipsilateral or contralateral DCIS or IBC (often a mixture)
We have limited data on DCIS progression with paired molecular profiles
The most consistent biological feature of DCIS:
Heterogeneity:
In clinical presentation
Morphology
Protein expression:
Including receptor status
Gene expression
Genetic alterations
Epigenetic alterations
The heterogeneity is:
Between patients – within the lesion – and within cells in a single duct
Morphological features that help us predict progression are:
Histologic grading:
Combing the nuclear grade 1 to 3 and necrosis into a three tier system (the good, the bad, and the ugly):
Low / intermediate / high grade
Grade 1 to 3
Van Nuys Group 1 to 3
DIN 1 to 3
We all know that there is regression towards the mean and substantial interobserver variation
DCIS is classified by two complementary systems:
Nuclear grade (the primary classification):
Low (grade I), intermediate (grade II), and high (grade III):
Based on nuclear size, pleomorphism, chromatin pattern, nucleoli, mitotic activity, and cell polarization
Nuclear grade is the most reproducible and clinically relevant parameter
Architectural pattern (secondary / noncore):
Solid, cribriform, micropapillary, papillary, and comedo:
Multiple patterns frequently coexist within a single lesion, limiting reproducibility
Morphologic Features Suggestive of Progression
Unclear prognostic value of the 3-tier system:
We suspect that:
Low to intermediate grade = low risk of progression
High grade system = high risk of progression or shorter time to progression
Maxwell, A.J. Eur.J.Surg.Oncol.,2018, Ryser, MD. J.Natl Cancer Inst., 2019:
Risk of ipsilateral recurrence (DCIS / IBC) at 10 years:
High grade 17.6% (95% CI=12.1-25.2%)
Non high grade 12.2 (95% CI=8.6-17.1%):
Including grade 2
There is overlap in the confidence intervals
Low grade DCIS are the lesions that might have:
Discontinuous growth (multifocal gaps up to 1 cm) and skip lesions that might lead to a:
Greater likelihood of residual disease and recurrence?
DCIS has heterogeneity of grade within the same lesion
Histologic subtype as a prognostic factor:
Subtype:
Cribriform:
Is more often a grade 1 lesion
Comedo type:
Is more often a grade 3 lesion
Usually histology subtype correlates with grade but:
There is often a mixture of growth patterns:
Compromising the use for prognostication
Can we use histology as a prognostic feature?
Tumor micro environment:
Could potentially be the most important morphologic feature suggestive of progression especially:
Circumferential periductal fibrosis and associated tumor infiltrating lymphocytes (TIL):
Indicating host reaction to the tumor cells
Tumor micro environment includes:
Myoepithelial cell layer
Tumor infiltrating lymphocytes (TIL)
Adipocytes
Fibroblasts
Matrix
The border around the myoepithelial layer might have prognostic value
The myoepithelial cell layer acts as a gatekeeper:
Has tumor suppressive functions
The largest gene expression change from normal tissue to DCIS:
Occurs in the myoepithelial cell layer
DCIS is associated myoepithelial cell loss:
That leads the decrease tumor suppressor functions
The myoepithelial cell layer is lost in IBC
Myoepithelial cell layer acts as a GATEKEEPER (Tumor Suppressive Functions)
Disruption of the myoepithelial defense:
IBC with DCISBrown color – myoepithelium Arrows point to the disrupted myoepithelial cell layer (micro-invasion)
Conflicting data on prognostic value of TIL:
Some studies have reported no prognostic value of stromal TIL for subsequent recurrences:
Does the spatial location of the immune cells matter?
The TIL in direct contact with the DCIS might be more important that the TIL that are further away
Other studies have shown a correlation between higher levels of TIL and increased risk of subsequent IBC and a shorter (ipsilateral) recurrence-free survival
The eighth edition of the American Joint Committee on Cancer (AJCC) staging system:
Defines microinvasion as:
Invasion of breast cancer cells:
Through the basement membrane at one or more foci:
None of which exceeds a dimension of 1 mm
DCIS:
Is a Tis lesion:
Is classified as stage 0 cancer
DCIS with microinvasion is considered:
T1mi:
Upstages DCIS from stage 0 to stage I disease:
The earliest stage of invasive cancer:
In the AJCC staging system
By definition:
DCIS does not have the ability to metastasize to axillary lymph nodes or distant sites:
Whereas DCIS with microinvasion does
Axillary metastasis:
Has been reported in 0% to 20% (0% to 28% in some series) of patients:
With DCIS with microinvasion
The incidence of microinvasion in DCIS:
Varies according to:
The size and extent of the index lesion
Lagios et al. (1989):
Reported a 2% incidence of microinvasion in patients with DCIS:
Measuring ≤ 25 mm in diameter
Compared with a 29% incidence of microinvasion:
In those with lesions ≥ than 26 mm
The incidence of microinvasion is also higher in patients with:
High-grade or comedo-type DCIS with necrosis
In patients with DCIS who present with:
A palpable mass
Nipple discharge
Historically, patients with DCIS with microinvasion:
Have been observed to have a worse prognosis:
Compared with those who have DCIS alone
Mirza et al. (2000):
Reported the long-term results of breast-conserving therapy in patients with:
DCIS
DCIS with micro-invasion (T1mi)
T1 invasive breast cancers
The 20-year disease-specific survival rates in patients with:
DCIS were better:
Than those among patients with DCIS with microinvasion or with T1 invasive tumors
Patients with micro-invasion (T1mi) and those with T1 tumors:
Had similar survival rates
In a retrospective study of 1,248 serially sectioned DCIS tumors, de Mascarel et al. (2002):
Reported a 10.1% incidence of axillary metastases:
In cases of DCIS with microinvasion
Patients with DCIS had a better 10-year distant metastasis-free survival rate:
Than patients with DCIS with microinvasion:
98% and 91%, respectively
The overall survival rate was also better in patients with DCIS compared to DCIS with microinvasion:
96.5% vs. 88.4%
However, the metastasis-free and overall survival rates:
Were worse in patients with invasive ductal carcinoma compared with those with DCIS with microinvasion
These results suggest that DCIS with microinvasion:
Should be characterized as a small invasive tumor with a good outcome:
The therapeutic approach for these patients should be similar to that for patients with invasive cancer
However, more recent studies have pointed toward DCIS with microinvasion having a more similar natural history to pure DCIS than to early-stage invasive disease:
In a review of 393 patients treated at Yale between 1973 and 2004:
There was no statistically significant difference between patients with DCIS and those with DCIS with microinvasion with regard to the presence of axillary metastases (in those who had axillary staging) or the likelihood of recurrence (locoregional and distant) or overall survival (Parikh et al., 2012)
Avoids all the disadvantages of a lateral mandibulotomy and the sequelae of a midline mandibulotomy
A paramedian mandibulotomy offers significant advantages:
Wide exposure to the surgical field
Preservation of the geniohyoid and genioglossus muscles:
Leading to preservation of:
The hyomandibular complex
The only muscle requiring division is the mylohyoid muscle:
Which leads to minimal swallowing difficulties
A paramedian mandibulotomy does not cause denervation or devascularization of the skin of the chin or the teeth and mandible
Fixation at the mandibulotomy site is easy
The site of the mandibulotomy does not fall within the lateral portal of radiation therapy if the patient needs postoperative radiotherapy:
Thus at present a paramedian mandibulotomy remains an optimal surgical approach for access to posteriorly located larger lesions of the oral cavity and tumors of the oropharynx and parapharyngeal space
The 2025 ATA guidelines (Recommendation 11) formally endorse active surveillance (AS) as an acceptable management option for cT1aN0M0 papillary thyroid cancers (≤1 cm), graded as a conditional recommendation with low certainty evidence.
For recurrent / persistent nodal disease, the guidelines state that stable disease can continue monitoring, while bulky and / or invasive recurrent disease is generally best treated with surgery.
Recommendation 14 specifies that surgical resection is indicated when there is tumor growth ≥ 3 mm, new biopsy-proven lymph node metastases, distant metastases, extrathyroidal extension, or patient preference.
The NCCN Thyroid Carcinoma Guidelines (v1.2026) similarly note that AS is supported by low-quality evidence for papillary thyroid microcarcinoma ≤ 1 cm, with limited data for cancers >1 cm.
Quantitative Data on Progression During Active Surveillance
A 2026 systematic review and meta-analysis (10 studies, n = 841) provides the most comprehensive pooled data. The overall pooled progression rate during AS was 23% (95% CI, 12–34%), rising to 32% in biopsy-confirmed cases. Critically, after adjustment for follow-up loss, the estimated progression rate increased to 35% to 70%, suggesting that reported rates may substantially underestimate true progression. Subgroup analyses showed higher progression in patients with elevated baseline thyroglobulin (log-transformed Tg: 1.02 in progressors vs. −0.07 in stable, p < 0.05) and in those with advanced stage (≥ 5% Stage IV patients: 28% vs. 14%, p < 0.05).
Individual cohort studies provide additional granularity:
Jerkovich et al. (2022): 50 patients with cytology-confirmed metastatic lymph nodes (mean size 10.1 mm) under AS; 24% had lymph node growth at median 29 months. The only predictor of growth was a Tg rise ≥0.5 ng/mL (OR 16.2, p = 0.020). Median progression-free survival was 6.6 years. Among the 7 patients who underwent surgery after progression, none had structural incomplete response at follow-up.
Walter et al. (2023): 40 patients with metastatic cervical lymph nodes ≤2 cm; 35% progressed (17.5% enlargement ≥3 mm, 17.5% new metastasis) at median 27.5 months. Of 14 who progressed, 6 remained on AS with 5/6 showing no further progression.
Tufano et al. (2015): Lateral neck lymph nodes with suspicious ultrasonographic features demonstrated a low potential for structural progression—over a median of 3.5 years, only 9% increased by more than 5 mm.
Size Thresholds: Confirmed
The ATA-endorsed size thresholds for FNA biopsy consideration (and by extension, the boundary for observation) are ≥8 mm in the central compartment and ≥10 mm in the lateral compartment. The NCCN guidelines do not specify exact size cutoffs for AS of recurrent nodes but recommend “disease monitoring for non-progressive disease that is stable and distant from critical structures”.
Thyroglobulin Doubling Time as a Prognostic Marker
A Tg doubling time (Tg-DT) <1 year portends poor prognosis, while Tg-DT >2 years signifies good prognosis. The AHNS consensus statement recommends cross-sectional imaging at 4 to 6 month intervals when Tg-DT is <1 year, with longer intervals for longer doubling times.
A study by Albano et al. demonstrated that a TgDT threshold of ≤2.5 years predicted positive FDG-PET/CT results with 93% sensitivity and 87% specificity (AUC = 0.911), outperforming absolute Tg level alone. Ito et al. (2025) showed that combining unstimulated Tg ≥3 ng/mL with a Tg doubling rate ≥0.33/year identified the highest-risk group for both lymph node and distant recurrence.
FDG-PET Avidity: Strong Prognostic Data
FDG-PET avidity is a well-established marker of aggressive biology. Robbins et al. demonstrated that a positive FDG-PET/CT predicted a 7-fold increased risk of mortality compared with negative FDG findings. FDG uptake reflects tumor dedifferentiation and is a major negative predictive factor for response to RAI treatment. An SUVmax >5.0 has been associated with very high rates of disease progression even in the presence of mixed iodine-avid disease, and SUVmax >10 has been linked to shorter locoregional disease-free survival and anaplastic transformation. The 2025 ATA guidelines recommend FDG-PET/CT as both a diagnostic and prognostic tool in high-risk patients, particularly those with aggressive histologies or negative RAI imaging.
BRAF/TERT Mutations: Quantified Risk
The co-occurrence of BRAF V600E and TERT promoter mutations is the most powerful molecular predictor of aggressive behavior. A meta-analysis showed this combination was associated with reduced disease-specific survival (RR 15.09, 95% CI 7.75–29.37), far exceeding BRAF V600E alone (RR 5.34). Song et al. (2016) reported that co-existing BRAF + TERT mutations increased recurrence risk (HR 4.64 for BRAF/TERT; HR 5.36 for RAS/TERT) and mortality (HR 15.13 for BRAF/TERT; HR 14.75 for RAS/TERT). Ryu et al. (2026) demonstrated that TERT-mutant tumors show continuous recurrence over time without the expected decline seen in TERT-wildtype cases, and are associated with significantly higher mortality. Xing et al. (2025) showed that the BRAF/TERT genetic duet worsened survival across all AJCC stages.
Reoperative Surgery Outcomes
The 2025 ATA guidelines report that reoperative surgery results in 27% to 63% excellent responses, 5.7% to 13.3% biochemically incomplete responses, 10% to 44% structurally incomplete responses, and 8.6% to 30% indeterminate responses. Independent risk factors for incomplete response after reoperation include age >45 years, aggressive histology, and lymph node ratio >0.6 at initial resection.
The Central Tension: Is Progression Evidence That Surveillance Is Failing?
The assertion that “progression itself is evidence that surveillance is failing” requires nuance. The data suggest that not all progression is equal:
Slow, limited nodal growth in low-risk patients can often be managed with continued observation or delayed intervention without compromising outcomes. In the Jerkovich cohort, all 7 patients who underwent surgery after progression on AS achieved structural complete response. In the Walter cohort, 5 of 6 patients who remained on AS despite progression showed no further growth.
However, the 2026 meta-analysis raises a cautionary note: true progression rates may be 2–3 times higher than reported due to follow-up loss, and AS “should be considered with caution”. Trimboli and Piccardo (2026) argue that until stronger prospective data are available, AS in DTC should be applied within a well-defined multidisciplinary framework with transparent communication about uncertainties.
The evidence therefore supports a nuanced position: AS is appropriate and evidence-based for carefully selected patients with stable, low-volume recurrence, but documented progression—particularly when accompanied by rising Tg kinetics, FDG avidity, aggressive molecular profile, or threatened critical structures—should prompt serious consideration of intervention. The key is that AS is not a passive strategy but an active decision-making framework requiring serial reassessment.
