Postoperative Surveillance of

Differentiated Thyroid

Carcinoma

Mitchell E. Tublin, M.D.

Vice Chair & Professor of Radiology

Section Chief: Abdominal Imaging

University of Pittsburgh

School of Medicine

Postoperative Surveillance of

DTC….and some observations

from a recovering cynic

Mitchell E. Tublin, M.D.

Vice Chair & Professor of Radiology

Section Chief: Abdominal Imaging

University of Pittsburgh

School of Medicine

Thyroid carcinoma: “epidemic”

• 2014 ACS estimate:

– 63K new cases (2009: 37K)

– 1,890 deaths, 0.3% of all cancer deaths (CA Cancer J Clin, 2014, Cancer Statistics, 2014)

• Majority are differentiated thyroid carcinoma

– Papillary cancer 80%

– Follicular cancer 11%

• Most slow growing and respond to tx

Thyroid carcinoma: recurrence basics

• Typically indolent tumor

• 5 year survival rate: 97.3%

• Tumor may present decades after initial presentation

• ↑ risk of recurrence, mortality if > 45, advanced tumor at initial presentation

Differentiated thyroid cancer

• Surveillance approaches for DTC rapidly

evolving in context of recent thyroid

carcinoma “epidemic”, improvements in

imaging, and molecular diagnostics

• ATA guidelines: 2006, 2009, 2015

Sono surveillance of DTC: My

world view

• I’m a sono advocate, I like scanning but I’m a

natural cynic

• Use of sono as primary modality for detection of

regional recurrence highlighted in recent (non-

radiology) literature

• Central role in 2015 ATA guidelines

• Imaging has outpaced understanding of biology

• Intervention often patient (and RVU) driven

DTC Surveillance:

controversies

• Which modality and when (it’s

still not all ultrasound…..)

• What criteria

• How report

• Should we care?

DTC surveillance: it’s all

about thyroglobulin

• Thyroglobulin

– Produced exclusively by follicular cells

– Should be undetectable after

thyroidectomy and successful ablation

– Elevated Tg indirect evidence of residual

functioning thyroid tissue or tumor

– Sensitivity of Tg 85-95% with L-T4

withdrawal or rhTSH stimulation

Thyroglobulin

• Very high negative predictive value

• False negatives may occur:

–Anti Tg Ab

–Absent or abnormal Tg production

by dedifferentiated metastases

131 (123)I Whole Body Scan

• Traditional strategy:

– Surveillance 131I whole body scan and

serum Tg post thyroidectomy and

ablation

• WBS utility discounted in 2009, 2015 ATA

guidelines

• Limited additional benefit if stimulated

thyroglobulin undetectable (and sono - )

Pacini, J Clin Endo & Metab, 2002

131 I Whole Body Scan

• 2009/2015 ATA guidelines

– Surveillance WBS not advocated for low

risk pts (no locoregional dz, favorable

histology) with neg Tg/sono

– Surveillance Dx WBS (1123 or low dose

I132) may be useful for intermediate or

high risk pts

131 I Whole Body Scan

• Still used if TG +, cervical US –

• With elevated Tg, 131I WBS localizes

recurrence in:

– 50 – 60% papillary carcinoma

– 64 – 67% follicular carcinoma

• Anatomic detail limited

131 I Whole Body Scan

• Potential causes for false negative WBS

– Insufficient TSH stimulation

– Iodine contamination

– Low volume tumor

– Dedifferentiated tumor

• Absence of 131I uptake in metastatic disease

may be a predictor of poor outcome

Fluorine-18-fluorodeoxyglucose

(FDG) PET

• Long accepted utility in Tg positive, WBS negthyroid cancer

– Sensitivity 60 - 94%

– Specificity 25 – 90%

• Recommended threshold Tg > 10 µg/L

• Increased FDG uptake and loss of 131I uptake may reflect dedifferentiation and more aggressive tumor (flip-flop phenomenon)

• Limited spatial resolution

• False positives (physiologic activity, inflammation)

Combined CT-PET

• Advantages

– Spatial localization

– Increased specificity (physiologic uptake)

• ATA 2015: FDG-PET / CT-PET synonymous

• Potential indications beyond Tg +, WBS – dz

– Initial staging & f/u of high risk pts with poorly differentiated dz

– Selection tool for pts unlikely to respond to 131I ablation

PET-CT Study Patients Sensitivity Specificity

Nahas et al. 33 66% 100%

Palmedo et al. 40 95% 91%

Iagaru et al. 76 88.6 89.3

Finkelstein et al. 65 98% 81%

Shammas et al. 61 68.4% 82.4%

Kim et al. 20 90%

Freudenberg et al. 36 96% 100%

UPitt (Ferris) 123 82.1% 89.7%

CT-PET: Literature

Thyroid cancerCT: 160 mAs; 130 KVp; pitch 1.6; 5 mm slices PET: 7 mCi FDG; 3 x 10 min; 3.4 mm slices

78 year-old male. Thyroid cancer

in 1993. Elevated thyroglobulin.

Negative 131I study. Four lesions identified on PET/CT

PET/CT scannerUniversity of Pittsburgh

131I scan

Thyroid cancer surveillance

• CT

– Useful for distant (lung) metastases

– Limited role for regional recurrence

• Size only useful criteria for nodal dz

• Anatomy may be obscured without IV contrast

• MRI

– Local dz, FSE T2, neck coil

– 1cm size threshold, high T2 signal (n = 26)

• Sensitivity 95% Specificity 51%

• PPV 84% NPV 78% Accuracy 83%

– Interobserver variability

Gross, Laryngoscope, 2001

42 y.o. female, pap ca, WBS neg, Tg +

FSET2

+Gd

Ultrasound

• Premier modality for thyroid imaging,

nodule identification, characterization (?)

and bx

• Optimum technique critical for imaging of

local-regional recurrence

– High frequency probe (10-14 MHz)

– Compound imagining

• Operative hx, zone classification

• Physician scanning crucial

Som, Curtin, Mancuso. Arch Otolaryngol-

Head Neck Surg, 1999

Ultrasound

• Combination of rhTSH-stimulated Tg and sonography most sensitive for monitoring differentiated thyroid ca regional recurrence in recent large series (n = 340)

Sens Spec

Sono 70 97.5131I WBS 20 100

Stim Tg 78.2 100

Sono/Tg 96.2 100

Pacini, J Clin Endocrinol Metab, 2003

Ultrasound: Lymph node

diagnositc criteria

• Grey scale

– Size

– Shape

– Internal architecture

• Color (power) Doppler

– Vascularity

– Pattern

Grey scale: malignant cervical

lymph nodes

• Size

– “large” (threshold dependent upon station)

• Short axis measurement

– Increased specificity with ↑ threshold

– ROC thresholds: 5mm (submental) → 9mm (submand)

Ying, JIUM,1998

• Clustering

• Shape

– Round (short axis/long axis ratio > .5)

Grey scale: malignant cervical

lymph nodes

• Internal architecture

– Hypoechoic

– Effaced echogenic hilum

• Unique features of papillary carcinoma

– Cystic

– Microcalcifications

41 y.o. female: “branchial

cleft cyst” at OSH

Metastatic papillary

carcinoma

Doppler: malignant cervical

lymph nodes

• Vascular pattern

– Capsular vessels

– “Low” resistive index (<.8)

– Issues

• Operator, parameter, dependent

• Interobserver variability

• Overlap

Non

necrotizing

granulomas

Suture

granulomas

Mimics

Mimics

Thoracic Duct Gel Foam

IJ

SC

Ultrasound: Lymph node

diagnostic criteria

• Combination of criteria improves sensitivity/specificity

– Ability of sono to depict internal grey scale and Doppler architecture increases accuracy relative to CT

– Comparison of sono and CT for differentiation of malignant and benign cervical lymph nodes in H/N SCCA (n = 62)

Sensitivity Specificity

CT .68 .82

Sono/doppler .78 .96Sumi, AJR 2001

• Ultrasound guided FNAC still crucial

Ultrasound: ATA 2015 triage

(recommendation # 65)

• Cervical ultrasound q 6-12 mos, and then “periodically” depending on Tgand risk level

• FNA of suspicious nodes > 8-10 mm short axis

• Tg aspirate recommended to ↑ yield

• “Cervical metastases may occasionally be detected at neck US even with undetectable Tg” (ATA 2009)

CS 211666020

• 28 y.o. female, total thyroidectomy, selective

bilateral neck dissection for multifocal, node

positive pap ca

• Completion ablation, right neck dissection 4/02

for recurrence

• Tg 226, neg WBS 8/02, 131I tx

• 2/03 Tg 73.5 / 10/03 Tg 100.5

• WBS neg, CT/PET: small left supraclav node,

mild thymic FDG uptake

• Ultrasound performed prior to resection

L4

L6

SM 163526819

• 30 y.o female, s/p thyroidectomy for tall cell variant papillary carcinoma 3 yrs prior

• 131I ablation x 2

• rhTSH 151

• WBS neg

• CT/PET

– 2 subcm right paratracheal nodules

– Diffuse left paratracheal uptake ? brown fat. “Correlation with Tg level and prior CT recommended.”

T

T

Central recurrence

R Level 6

R Level 4

Compound & SR14 MHz

MP 050446785

• 45 y.o male, post thyroidectomy, Left RND

for papillary carcinoma

• 131I ablation

• Persistent elevated Tg despite selective

node dissection WBS: mild uptake within

thyroid bed

• MRI: small left level 6 node

• Level 6 dissection (metastatic pap ca)

T

L Level 6

L4R4

“Micrometastases”

More “micrometastases”: 3 patients

Differentiated thyroid cancer

• Subclinical nodal metastases probably present

in high percentage of patients with DTC

– No value of prophylactic lymph node

dissection in 57 patients despite high

incidence (90%) of nodal metastases

– 55% of nodal metastases < 5mm (Noguchi,

Cancer, 1970)

Thyroid cancer incidence and mortality, 1973-2002

Davies, Welch, JAMA, 2006

Thyroid CA:

trends

Type

Size

Davies, Welch, JAMA, 2006

And it’s accelerating: Morris, Sikora, Tosteson, Davies. The increasing incidence of thyroid cancer:

the influence of access to care. Thyroid, Nov 2013.

Thyroid-Cancer Incidence and Related Mortality in South Korea,

1993–2011

Ahn HS et al. N Engl J Med 2014;371:1765-1767.

• 50 fold increase incidence, 1993-2011

• Most common CA dx (> 40K, 2011)

Penetration of Thyroid-Cancer Screening (2008–2009) and Incidence of

Thyroid Cancer (2009) in the 16 Administrative Regions of South Korea

Ahn HS et al. N Engl J Med 2014;371:1765-1767.

Ultrasound surveillance of DTC:

practical impact

• Increased incidence of thyroid carcinoma and improved detection of recurrent or residualdisease largely due to ultrasound

• Aggressive surveillance advocated – yearly Tg and neck ultrasound

• Repeated nodal dissections for low volume locoregional recurrence/residual often performed

– Patient anxiety

– $$$$$

Ultrasound surveillance of DTC:

practical impact

• ? Improved survival with aggressive

surveillance and tx of typically indolent tumor

• The cynic:

– “First, do no harm”

– “Don’t ask, don’t tell”

– “What you don’t know won’t hurt you”

– “Turn the machines off”

Cronan, “Thyroid Nodules: Is it time to

turn of the US machines?” Radiology,

2008

34 y.o. female: 3.6 cm tall cell variant, post

thyroidectomy/ablation, 3 yrs prior. Tg 19.

FNA: 7 mm L6 pap ca nodal recurrence.

Difficult resection despite sono localization.

The real (but unspoken) debate

Should we do any postoperative surveillance at all????

Answer: yes, but…

Ultrasound surveillance of DTC:

practical impact

• Biologic variability

• Small subset of even “nano” papillary

carcinoma may be aggressive

Kuffner : Microfoci of papillary carcinoma are

frequent and can be metastatic. Retrospective

analysis of a cohort of patients from the University

of Pittsburgh Thyroid Cancer Database

32 y.o. female, 6mm

micropapillary carcinoma.

Right neck dissection.

45 y.o. ♂, CT for

“night sweats”:

L 4 node

2 mm micro BRAF +

PTC, 3 cm L4 node

Surveillance of DTC: observations from a

recovering cynic

• We’re learning and adapting

• Machines still on, but improved pathology lexicon, molecular diagnostics, risk stratification, radiology/endocrinology decision trees must be making a difference

• I’m seeing it locally

UPMC Multidisciplinary Thyroid Clinic

• Paradox

• RVU centric: intervention driven

• Marketing bonanza

• Broad Western PA referral base

• Volume originally not an issue

• Stated goal: rational utilization of resources for indolent dz

UPMC Multidisciplinary Thyroid Clinic

• High volume practice early on

• A great clinical model: expertise leveraged

• Victim of our own success

• Tailored clinical pathways, built upon high impact,

collaborative research

• Volume down….a lot• 43% ↓ total FNAs since 2010 (1400 – 800)

• Clinic largely dissolved: many of original core

have left

• Radiology (me) no longer physically on site: tele-

sonography

De-escalating management of patients

with thyroid nodules and cancer

• Fewer FNAs (ATA, and increasingly ACR

TIRADS)

• Molecular diagnostics (Thryoseq 1,2,3)

• Fewer diagnostic lobectomies

• Fewer completion thyroidectomy

• More conservative staging

• Active surveillance for small PTC

• Cancer nomenclature revision

• Encapsulated follicular variant of PTC → Non-

invasive follicular thyroid neoplasm with

papillary-like nuclear features (NIFTP)

2015 ATA Risk of Structural Disease Recurrence

Low Risk

Intermediate Risk

High Risk

FTC, extensive vascular invasion (≈ 30-55%)

pT4a gross ETE (≈ 30-40%)

pN1, any LN > 3 cm (≈ 30%)

BRAF mutated, not intrathyroidal (≈ 10-40%)

PTC, vascular invasion (≈ 15-30%)

Clinical N1 (≈20%)

pN1, > 5 LN involved (≈20%)

BRAF mutated, intrathyroidal, < 4 cm (≈10%)

pT3 minor ETE (≈ 3-8%)

pN1, all LN < 0.2 cm (≈5%)

pN1, < 5 LN involved (≈5%)

Intrathyroidal 2-4 cm PTC (≈ 5%)

Multifocal PMC (≈ 4-6%)

Minimally invasive FTC (≈ 2-3%)

BRAF wild type, intrathyroidal, < 4 cm (≈ 1-2%)

BRAF mutated, intrathyroidal unifocal PMC (≈ 1%)

Intrathyroidal, encapsulated, FV-PTC (≈ 1-2%)

Unifocal PMC (≈ 1-2%)Slide courtesy of

Steven Hodak, MD

Molecular Signature

BRAF+TERT, RAS+TERT

Multiple driver mutations

(eg. NRAS and PIK3CA or TP53)

TERT

ALK fusions

NTRK1 fusions

NTRK3 fusions

BRAF V600E

RET/PTC

RAS

BRAF K601E

PAX8/PPARG

RAS-like

mutations

BRAF

V600E- like

mutations

Molecular markers for cancer risk stratification

High RiskGross extrathyroidal extension,

incomplete tumor resection, distant metastases,

or lymph node >3cm

Intermediate RiskAggressive histology, minor extrathyroidal extension,

vascular invasion,

or > 5 involved lymph nodes (0.2-3 cm)

Low RiskIntrathyroidal DTC

≤ 5 LN micrometastases (< 0.2 cm)

Risk of Structural Disease Recurrence

(In patients without structurally identifiable disease after initial therapy)

PTC, >1 cm, TERT mutated ± BRAF mutated* (>40%)

PTC, extrathyroidal, BRAF mutated* (≈ 10-40%)

Intrathyroidal PTC, < 4 cm, BRAF mutated* (≈10%)

Intrathyroidal, < 4 cm, BRAF wild type* (≈ 1-2%)

Intrathyroidal unifocal PTMC, BRAF mutated*, (≈ 1-2%)

FTC, extensive vascular invasion (≈ 30-55%)

pT4a gross ETE (≈ 30-40%)

pN1 with extranodal extension, >3 LN involved (≈ 40%)

pN1, any LN > 3 cm (≈ 30%)

PTC, vascular invasion (≈ 15-30%)

Clinical N1 (≈20%)

pN1, > 5 LN involved (≈20%)

pT3 minor ETE (≈ 3-8%)

pN1, all LN < 0.2 cm (≈5%)

pN1, ≤ 5 LN involved (≈5%)

Intrathyroidal PTC, 2-4 cm (≈ 5%)

Multifocal PMC (≈ 4-6%)

pN1 without extranodal extension, ≤ 3 LN involved (2%)

Minimally invasive FTC (≈ 2-3%)

Unifocal PMC (≈ 1-2%)

Intrathyroidal, encapsulated, FV-PTC (≈1-2%)

NIFTP (<1%)

Haugen BR et al. Thyroid. 2016, 26:1-133

❑ A distinct class of thyroid tumors:• Clonal process driven by distinct oncogenic mutations (RAS and

RAS-like gene mutations) • Non-invasive, follicular-patterned, moderately to well developed

nuclear features of PTC (nuclear score 2-3)• Highly favorable outcome (<1% risk of recurrence in 15 y)

❑ Recommended new terminology:

“Non-Invasive Follicular Thyroid neoplasm with Papillary-like nuclear features“ (NIFTP)

Nikiforov et al. JAMA Oncology 2016; 2:1023-9.

Molecular Markers Predict Cancer Risk

n=1510

1. Yip L Ann Surg 2015; 262: 519

2. Song YS Cancer 2016; 122: 1370

p<0.001

DTC, n=1510

High Risk for Recurrence2Markers of Low and

Intermediate Recurrence

Risk1

DTC, n=551

p<0.001

Bethesda III-IV

Cytology

N/A

Observation Lobectomy

Test result

Probability of Cancer or NIFTP

Tumor type, risk of recurrence

Patient management

3-4% 40-80%

NIFTP or low-risk cancer

Active surveillance

<10%

ThyroSeq GC

Positive: RAS-like

mutations,CNA, GEA

Negative: no mutations

Currently Negative:

LR mutations

NIFTP or low-risk cancer

Total thyroidor lobectomy

95-100%

Positive: BRAF-like mutations

Intermediate-risk cancer

Total thyroid+/- LND

98-100%

Positive: HR mutation

High-risk cancer

MTC, PT, Non-FCL

28% 6% 2%7%55%

2%

Nikiforov YE et al. Real world experience with ThyroSeq v3 Genomic Classifier in thyroid nodules with indeterminate

cytology: Results of testing of the first 3,734 consecutive samples. American Association of Clinical Endocrinologists

Annual meeting 2018

Patient Management Informed by Molecular Testing

2014

2016

2017

2018

Conclusions

• Increased incidence of thyroid carcinoma over

past thirty years largely due to imaging

• DTC usually indolent dz, but local recurrence or

detection of previously unrecognized tumor may

occur long after initial presentation

• TG and ultrasound currently best combination for

regional surveillance

• Sensitivity/specificity of sono improved when

multiple discriminatory criteria used; FNAC still

necessary

Conclusions

• Impact of intensive surveillance/tx difficult

to assess…but it’s probably not good

• Improvements in imaging have to now

outpaced our understanding of DTC tumor

biology

• More rational approach needed, and it will

be driven by recent advances in molecular

diagnostics (and insurers)

Surveillance of DTC: future

• New guidelines for surveillance intensity likely given better understanding of tumor biology

• Expect less surveillance ultrasound volume

• Healthy cynicism can be a good thing