ThyroidCA
Differentiated Thyroid Cancer · Clinical Decision Support · ATA 2025
A product of EndoHormone Institute (EHI)
v1.3.0
Complete patient data on the left and click Generate Clinical Plan to produce an ATA 2025-aligned management recommendation. Cases can be saved to the database for later retrieval.
| Driver | First-Line Agent(s) | Notes |
|---|---|---|
| RET fusion | SelpercatinibPralsetinib | Highly selective; superior RR vs MKIs |
| NTRK1/3 | LarotrectinibEntrectinib | Tumour-agnostic approvals |
| BRAF V600E | Dabrafenib + Trametinib | Consider redifferentiation before RAI |
| ALK fusion | Alectinib | Emerging data; enrol in trial if available |
| No driver | LenvatinibSorafenib | MKI first-line; TMB-H/MSI-H (rare) consider ICI |
ATA 2025 supports intent-based RAI selection. Individualise based on risk features, disease burden, and MDT nuclear medicine input.
| Subtype | ATA 2025 Risk | Key Diagnostic Features |
|---|---|---|
| Tall Cell Variant PTC | Min. Intermediate | Height ≥2x width, >30% of tumour; BRAF V600E common; reduced RAI avidity |
| Columnar Cell Variant PTC | High | Nuclear stratification; aggressive; higher disease-specific mortality |
| Hobnail / Micropapillary PTC | Min. Intermediate-High | >30% hobnail features; loss of polarity; BRAF V600E; higher recurrence & mortality |
| Diffuse Sclerosing Variant PTC | Min. Intermediate | Bilateral diffuse involvement; dense fibrosis; psammoma bodies; frequent LN metastases |
| Solid / Trabecular Variant | Min. Intermediate | >30% solid growth; exclude PDTC morphology (Turin criteria) |
| PDTC | High | Turin criteria: solid/insular growth + mitoses ≥3/10 HPF, necrosis, or convoluted nuclei |
| OTC High-Grade Features | High | Mitoses ≥3/10 HPF or tumour necrosis; poor RAI avidity |
Lowest baseline aggressiveness. Risk driven by extent, nodal burden, and ETE. RAI avidity usually preserved. Vascular invasion is a weaker prognostic driver than in follicular-pattern tumours.
Encapsulated disease is indolent; infiltrative disease carries higher recurrence risk. RAI generally preserved but can rise with infiltration or significant vascular invasion.
Vascular invasion burden is the dominant prognostic driver. Haematogenous spread more common; nodal disease uncommon. RAI variable. <4 foci = lower/intermediate; ≥4 foci = high risk.
Higher baseline aggressiveness than FTC. RAI avidity lower and less predictable. Any vascular invasion escalates risk; ≥4 foci treated as high-risk biology.
The ATA 2025 guidelines recognise several histological subtypes of DTC that carry intrinsically higher recurrence risk and/or poorer prognosis than classical PTC. Selecting one of these automatically escalates the ATA recurrence risk category.
Tall cells defined as height ≥2x width. BRAF V600E highly prevalent (~80%). Carries higher recurrence and disease-specific mortality vs classical PTC. Cannot be classified as Low Risk. Associated with reduced RAI avidity and BRAF-mediated dedifferentiation.
Rare, aggressive variant. Nuclear stratification is the hallmark. High recurrence and mortality rates. Total thyroidectomy and adjuvant RAI strongly recommended. Classify as High Risk.
≥30% hobnail features (loss of cellular polarity, prominent nucleoli, micropapillary architecture). BRAF V600E common. Associated with higher recurrence and disease-related mortality. Classify as at least Intermediate-High.
Bilateral diffuse thyroid involvement, dense fibrosis, psammoma bodies, squamous metaplasia, lymphocytic infiltration, frequent lymph node metastases. At least Intermediate risk; total thyroidectomy required.
>30% solid/trabecular growth in otherwise DTC morphology. Must be distinguished from PDTC (apply Turin criteria). Associated with higher recurrence. At least Intermediate risk.
Defined by Turin criteria: solid/trabecular/insular growth PLUS at least one of: convoluted nuclei, mitoses ≥3/10 HPF, or coagulative tumour necrosis. Limited RAI responsiveness; early systemic therapy consideration. High Risk by definition.
Oncocytic carcinoma with mitoses ≥3/10 HPF or tumour necrosis. Classified separately from standard OTC under WHO 2022 classification. Poor RAI avidity; High Risk. Systemic therapy consideration early.
Multifocal DTC (two or more distinct tumour foci) is common in PTC. ATA 2025 provides the following guidance:
Standard risk stratification applies based on size, extent, nodes, and margins alone.
Ipsilateral multifocality alone is generally not a significant independent risk escalator for intrathyroidal disease. Risk is driven by the dominant (largest) focus and aggregate adverse features.
Bilateral multifocality and/or ≥3 foci may upgrade recurrence risk, particularly when combined with other intermediate- or high-risk features (e.g. ETE, large nodes, unfavourable histology). This system applies a risk increment for bilateral or extensively multifocal disease.
Bilateral multifocal PTC is one of the strongest indications for total thyroidectomy over hemithyroidectomy. It ensures complete excision, enables Tg surveillance, and facilitates RAI if required. Ipsilateral multifocality of ≤2 foci does not automatically require total thyroidectomy if other features support lobectomy.
ATA 2025 guidelines support intent-based selection of RAI activity rather than fixed dosing. Individualise based on intent, pathological risk features, disease burden, and MDT nuclear medicine input.
Elimination of normal thyroid remnant post-surgery. For low-risk and selected intermediate-risk patients. Comparable efficacy to higher activities with reduced toxicity.
Reduction of recurrence risk with adverse pathological features. Context-dependent; integrated risk assessment not histology alone.
Higher activities; empiric or dosimetry-guided. Especially for distant metastases or repeat treatment. Nuclear medicine-led; diagnostic WBS recommended first.
ThyroidCA is a browser-based clinical decision support tool for the multidisciplinary management of differentiated thyroid carcinoma (DTC), developed and maintained by Professor Steven C. Boyages at the EndoHormone Institute (EHI), Sydney. The tool synthesises the 2025 American Thyroid Association (ATA) guidelines to provide structured, individually tailored management recommendations spanning surgical planning, radioactive iodine therapy, TSH suppression, systemic and targeted therapy, serum Tg interpretation, and post-treatment surveillance.
The tool was developed to address a recognised gap in accessible, guideline-aligned decision support for DTC management in clinical practice. While the ATA 2025 guidelines are comprehensive, their application to individual patients — with their specific combinations of tumour, pathological, and molecular features — requires integration of multiple interacting variables. ThyroidCA operationalises this integration in real time, providing a structured aide-memoire for clinical encounters, MDT discussions, and post-operative planning.
The AI pathology auto-fill feature uses the Anthropic Claude API to extract structured clinical data directly from uploaded surgical pathology reports (PDF or image), reducing manual data entry and transcription error. Documents are processed via an encrypted server-side proxy; no patient data is stored by EHI.
ThyroidCA implements a weighted scoring algorithm aligned to the ATA 2025 recurrence risk stratification framework. Each clinical and pathological feature is assigned a score based on its established prognostic weight in the guidelines:
Risk tiers: Low (score <12), Intermediate (12–34), High (≥35). Certain features (M1 disease, PDTC, columnar cell carcinoma) trigger automatic High-risk classification regardless of score. The surgical procedure type modifies downstream recommendations — particularly RAI eligibility and Tg surveillance — without changing the tumour risk tier itself.
ThyroidCA is designed for use by qualified healthcare professionals involved in the management of thyroid cancer, including endocrinologists, endocrine surgeons, nuclear medicine physicians, medical oncologists, radiation oncologists, and clinical nurse specialists. It is intended as a structured aide for clinical decision-making, MDT preparation, and post-operative planning — not as a standalone diagnostic or prescriptive tool.
Optimal use involves entering pathological data from the surgical pathology report, adding available molecular and biochemical data, and reviewing the generated plan in the context of the individual patient's clinical circumstances, comorbidities, and preferences. The AI auto-fill feature can accelerate data entry from pathology reports but should always be reviewed for accuracy before a plan is generated.
ThyroidCA provides structured guideline-based recommendations and cannot account for all factors relevant to individual clinical decisions. Known limitations include:
Planned features under active development at EHI include: integration of ATA response-to-therapy reassessment categories (Excellent / Indeterminate / Biochemical Incomplete / Structural Incomplete) with dynamic follow-up plan generation; TNM staging calculator (AJCC 8th edition); paediatric DTC module; anaplastic and medullary thyroid carcinoma decision support; export to structured PDF report; and integration with Australian PBS/TGA prescribing information for targeted therapies. Guideline updates will be incorporated as new ATA recommendations are published.
Clinical errors, suggested improvements, and feedback are welcomed. Please contact Professor Boyages directly at steven.boyages@sydney.edu.au or through the EndoHormone Institute at endohormoneinstitute.com.au. Reports of clinical inaccuracies will be reviewed and addressed in the next version update.
ThyroidCA was developed entirely by Professor Boyages and the EndoHormone Institute using independent resources. No pharmaceutical, device, or industry funding has been received for the development or maintenance of this tool. Professor Boyages has received honoraria for educational presentations from Sanofi, Novo Nordisk, and Eli Lilly in the fields of diabetes and obesity; no industry relationships exist in thyroid oncology. The tool does not promote any specific commercial product, diagnostic service, or institution.
This tool is intended solely as an educational aid and clinical decision support resource. It does not replace clinical judgement, multidisciplinary team (MDT) discussion, institutional protocols, or the doctor–patient relationship. All clinical decisions remain the sole responsibility of the treating clinician.
Designed exclusively for qualified healthcare professionals. Must not be used by patients, members of the public, or those without appropriate clinical training to interpret its outputs.
Patient data is processed locally in your browser. Pathology documents uploaded for AI extraction are transmitted securely via an encrypted server-side proxy to Anthropic's API and are not stored by EHI or Anthropic beyond the API call. Patient names are optional. Do not upload identifiable documents without appropriate consent and institutional privacy compliance.
Recommendations are based on: American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer (ATA 2025 revision). Users are directed to the original ATA publication for full guideline text, evidence grading, and supporting references.
📓 Open User Manual & Help Directory — full field guide, AI extraction guide, subscription help, and troubleshooting FAQ.
ⓘ Independent tool — not affiliated with the ATA. ThyroidCA is an independent clinical decision support resource developed by EndoHormone Institute Pty Ltd (EHI). It is not affiliated with, endorsed by, approved by, or certified by the American Thyroid Association® (ATA®). References to ATA 2025 guidelines are factual descriptors indicating the guideline version upon which recommendations are based. AMERICAN THYROID ASSOCIATION® and ATA® are registered trademarks of the American Thyroid Association, Inc.
© 2025–2026 EndoHormone Institute Pty Ltd (EHI). All rights reserved. ThyroidCA is a product of EHI. Unauthorised reproduction or redistribution is prohibited. For licensing enquiries contact steve.boyages@gmail.com.