Follicular Variant of Papillary Thyroid CarcinomaA Clinicopathologic Study of a Problematic Entity

Jeffrey Liu, MD1

Bhuvanesh Singh, MD, PhD2

Giovanni Tallini, MD3

Diane L. Carlson, MD4

Nora Katabi, MD4

Ashok Shaha, MD2

R. Michael Tuttle, MD5

Ronald A. Ghossein, MD4

1 Department of Otolaryngology-Head and NeckSurgery, Columbia University, New York, New York.

2 Department of Surgery, Memorial Sloan-Ketter-ing Cancer Center, New York, New York.

3 Department of Pathology, Hospedale Bellaria,Bologna, Italy.

4 Department of Pathology, Memorial Sloan-Ket-tering Cancer Center, New York, New York.

5 Department of Medicine, Endocrinology Service,Memorial Sloan-Kettering Cancer Center, New York,New York.

BACKGROUND. There is continuous debate regarding the optimal classification,

prognosis, and treatment of the follicular variant of papillary thyroid carcinoma

(FVPTC). The objective of this study was to assess the behavior of FVPTC, espe-

cially its encapsulated form, and shed more light on its true position in the clas-

sification scheme of well differentiated thyroid carcinoma.

METHODS. All patients with FVPTC, follicular thyroid adenoma (FTA), and fol-

licular thyroid carcinoma (FTC) who were diagnosed between 1980 and 1995

were reviewed and reclassified according to the currently accepted definition

of FVPTC. The tumors were separated into encapsulated and nonencapsulated

(infiltrative/diffuse) types. Encapsulated tumors were subdivided further into

tumors with or without capsular/vascular invasion. These different subtypes

of FVPTC were correlated with outcome and with other clinicopathologic pa-

rameters.

RESULTS. After review by 4 pathologists, 78 patients were included in the study.

Sixty-one of 78 patients (78%) had encapsulated tumors (18 invasive, 43 noninva-

sive), and 17 patients had nonencapsulated tumors (infiltrative/diffuse). The gen-

der distribution, age at presentation, and tumor size did not differ between

patients with encapsulated and nonencapsulated FVPTC. Patients who had encap-

sulated FVPTC had a significantly lower rate of marked intratumor fibrosis (18%),

extrathyroid extension (5%), and positive margins (2%) compared with patients

who had nonencapsulated tumors (88%, 65%, and 50% respectively; P < .0001).

Regional lymph node metastases were present in 14 of 78 patients (18%), and no

patients had distant metastases. The lymph node metastatic rate was significantly

higher in patients who had nonencapsulated tumors (11 of 17 patients; 65%) com-

pared with patients who had encapsulated neoplasms (3 of 61 patients; 5%; P <

.0001). In addition, lymph node metastases were not detected in any noninvasive,

encapsulated FVPTCs. With a median follow-up of 10.8 years, only 1 patient

developed a recurrence, which occurred in an encapsulated FVPTC that had

numerous invasive foci. None of the patients with noninvasive, encapsulated

FVPTCs developed recurrences, including 31 patients who underwent lobectomy

alone, with a median follow-up of 11.1 years.

CONCLUSIONS. FVPTC appeared to be a heterogeneous disease composed of 2

distinct groups: an infiltrative/diffuse (nonencapsulated) subvariant, which re-

sembles classic papillary carcinoma in its metastatic lymph node pattern and

invasive growth, and an encapsulated form, which behaves more like FTA/FTC.

Patients who had noninvasive, encapsulated FVPTCs did not develop lymph node

metastases or recurrences and could be treated by lobectomy alone. If the cur-

rent findings are confirmed, then strong consideration should be given to reclas-

sifying encapsulated FVPTC as an entity that is close to the FTA/FTC class of

tumors. Cancer 2006;107:1255–64. � 2006 American Cancer Society.

KEYWORDS: follicular variant, papillary, thyroid, carcinomas, infiltrative, diffuse,encapsulated.

Address for reprints: Ronald A. Ghossein, MD,Department of Pathology, Memorial Sloan-KetteringCancer Center, 1275 York Avenue, New York, NY10021; Fax: (212) 717-3203; E-mail: ghosseir@mskc.org

Received April 13, 2006; revision received June3, 2006; accepted June 13, 2006.

ª 2006 American Cancer SocietyDOI 10.1002/cncr.22138Published online 9 August 2006 in Wiley InterScience (www.interscience.wiley.com).

1255

W ell differentiated thyroid tumors of follicular cell

origin are subdivided into follicular thyroid ad-

enoma (FTA), follicular thyroid carcinoma (FTC), and

papillary thyroid carcinoma (PTC). PTC is the most

frequent type of thyroid malignancy (>70% of

tumors): It often is multifocal, nonencapsulated, and

spreads through the lymph nodes.1 In contrast, FTC

is unifocal, encapsulated, and metastasizes directly

to distant organs with a very low incidence of lymph

node metastases.1 The nuclear features of papillary

carcinoma (e.g., nuclear clearing, grooves, pseudoin-

clusions) characterize a carcinoma that belongs to

the PTC family.1 The follicular variant of PTC

(FVPTC) is the most common subset of papillary car-

cinoma and is found in 9% to 22.5% of patients with

PTC.2–5 This variant is composed entirely or almost

completely of follicles, which are lined by cells that

have the nuclear features of papillary carcinoma.6

Thus, FVPTC shares with FTA and FTC the presence

of follicles. When FVPTC is nonencapsulated and

infiltrates the surrounding thyroid parenchyma or

diffusely involves the thyroid, the diagnosis of carci-

noma usually poses no problem.7 For the encapsu-

lated tumor without invasion of surrounding thyroid

tissue, the diagnosis of malignancy relies solely on

the presence of the nuclear features of PTC (e.g., nu-

clear clearing, grooves, pseudoinclusions), which of-

ten can be borderline. Therefore, the diagnosis of

noninvasive, encapsulated FVPTC versus follicular

adenoma is prone to considerable interobserver vari-

ability.7,8 This diagnostic dilemma has very important

therapeutic implications. Indeed, if an FVPTC meas-

ures �1.5 cm, then many physicians in the U.S. will

recommend completion thyroidectomy followed by

radioactive iodine therapy (RAI).9 Some authors have

suggested that patients with encapsulated, noninva-

sive FVPTC have an excellent prognosis and, thus,

believe that only a lobectomy is needed. However,

there are no outcome data with long median follow-

up from a large number of patients with FVPTC.

More important, to our knowledge, there has been

no study in which tumor behavior was analyzed ac-

cording the histologic ‘‘subvariants’’ of FVPTC. (i.e.,

nonencapsulated [infiltrative/diffuse] vs. encapsulated)

that can serve as the basis for a conservative treat-

ment approach of encapsulated, noninvasive FVPTC.

In addition, there is some controversy regarding the

classification of FVPTC as a member of the PTC

group versus the FTA/FTC group. Indeed, Baloch and

LiVolsi showed that some encapsulated FVPTCs

metastasize to distant sites in the absence of lymph

node metastases, mimicking the behavior of FTC.10

Other authors reported that FVPTC has a signifi-

cantly lower metastatic lymph node rate and more

often is encapsulated than classic PTC.11,12 Recently,

several groups have attempted to analyze FVPTC at

the molecular and chromosomal levels.13,14 All of

those studies concurred that the molecular profile of

the FVPTC seems to be closer to the FTA/FTC group

than to classic PTC, supporting further consideration

of the classification of FVPTC.

To assess the behavior of FVPTC (especially its

encapsulated form) and to shed more light on its

true position in the classification scheme of well dif-

ferentiated thyroid carcinoma, we undertook a clini-

copathologic study of all patients with FVPTC who

were seen at Memorial Sloan-Kettering Cancer Cen-

ter between 1980 and 1995. FVPTC tumors were clas-

sified according to histologic growth patterns in

encapsulated versus nonencapsulated (infiltrative/

diffuse) neoplasms. The encapsulated subset was

subdivided further according to the presence or ab-

sence of invasion in a manner similar to that used to

differentiate FTA from FTC.

MATERIALS AND METHODSInclusion CriteriaThe Memorial Sloan-Kettering Cancer Center pathol-

ogy data base was searched for all tumors that had a

diagnosis of FVPTC, FTA, and FTC between January

1980 and December 1995. All tumors with adequate

material captured by the computerized data base

search were examined microscopically by a head and

neck surgical pathologist with expertise in thyroid

neoplasia (R.A.G.). Tumors were classified as FVPTC

if they were composed completely or almost entirely

(99% of the tumor) of follicles lined by cells that had

the nuclear features of PTC (i.e., irregular, enlarged,

clear nuclei with grooves, pseudoinclusions, and over-

lapping).6 The presence of tumor necrosis or increased

mitotic activity (�5 per 10 high-power fields) excluded

the tumor from the FVPTC category.

The following FVPTCs were included in the study:

FVPTC that measured �1 cm in greatest dimension

with no associated thyroid malignancy except for the

presence of 1 or 2 subcentimeter papillary microcar-

cinomas. For example, FVPTC associated with 3 foci

of papillary microcarcinomas were excluded along

with FVPTC associated with any non-FVPTC that mea-

sured �1 cm in greatest dimension. FVPTC asso-

ciated with any nonpapillary thyroid malignancy also

was excluded.

Pathology ReviewExcept for one tumor (reviewed by R.A.G. only), all

tumors that were included after the initial review were

reexamined microscopically by four board-certified

1256 CANCER September 15, 2006 / Volume 107 / Number 6

pathologists with special interest in thyroid neopla-

sia, (R.A.G., G.T., D.L.C., N.K.) to confirm the diagno-

sis of FVPTC. Histopathologic review was conducted

without the pathologist’s knowledge of the patients’

clinical characteristics or outcome. FVPTCs were sub-

divided into the following histologic subvariants: 1)

encapsulated if the tumor was surrounded totally by

a tumor capsule with or without capsular/vascular

invasion; 2) infiltrative if there was absent or incom-

plete encapsulation with invasive tongues of tumor

infiltrating nonneoplastic thyroid parenchyma, almost

always with prominent fibrosis; or 3) diffuse if 1 lobe

or an entire thyroid was involved by a nonencapsu-

lated, diffuse, or multinodular tumor without desmo-

plasia and with pushing borders or absence of a

clear cut delineation between the tumor and the ad-

jacent parenchyma. In view of their lack of total

encapsulation, the infiltrative and diffuse variants

both were categorized as nonencapsulated. The dis-

tribution of the nuclear features characteristic of

papillary carcinomas was recorded as either diffuse

(i.e., throughout the lesion) or multifocal (i.e., foci

that contained papillary carcinoma nuclei were dis-

tributed randomly over 30% of the tumor intermixed

with areas where the nuclear features were not devel-

oped fully or were absent). Tumor size was measured

as the greatest dimension of the resected tumor speci-

men. Mitotic rate was determined by counting 10

high-power fields (�400 magnification) with an Olym-

pus microscope (U-DO model) in the areas of greatest

concentrations of mitotic figures. Vascular and capsu-

lar invasion were identified according to the criteria

outlined in the last Armed Forces Institute of Pathol-

ogy fascicle on thyroid tumors.6 Capsular invasion was

defined as complete penetration of the entire thick-

ness of the tumor capsule. Irregularities of the contour

along the inner border of the capsule or nests of tu-

mor embedded within the capsule were not consid-

ered capsular invasion. Vascular invasion was defined

as invasion of a vessel located within or outside the

tumor capsule. The intravascular tumor growth had to

be covered by endothelium or attached to the vessel

wall. Irregular clusters of nonendothelialized tumor

cells that were not attached to the vessel wall repre-

sented artefactual dislodgement of tumor during sec-

tioning and did not qualify as vascular invasion. The

numbers of foci of vascular invasion in tumor capsule

(intracapsular vascular invasion) and outside the tu-

mor capsule (extracapsular vascular invasion) were

recorded. The numbers of foci of capsular invasion

also were counted. The cytoplasm of the predominant

cell type was categorized as either oncocytic or nonon-

cocyctic. The degree of intratumoral fibrosis was

recorded as absent, mild, or marked. The presence or

absence of tumor extension into the extrathyroid soft

tissue stroma as well as the presence of extrathyroid

vascular invasion were documented. Finally, micro-

scopic resection margins were categorized as either

positive (tumor at the inked margin) or negative (no

tumor at the inked margin).

Clinical ReviewThe patient’s charts were reviewed for age at diagno-

sis, gender, type of surgery, and adjuvant therapy (ra-

dioactive iodine). Tumor recurrence, both local and

metastatic, was established on the basis of clinical

examination; radiologic findings on either computer

tomography or magnetic resonance imaging studies;

findings on follow-up nuclear scanning, including

RAI or positron emission tomography scanning; sig-

nificant rise in serum thyroglobulin; or histologic ex-

amination of the recurrent tumor. Dates of initial

thyroid surgery, last follow-up, recurrence, and death

were recorded. Status at last follow-up was categor-

ized as no evidence of disease, alive with disease,

dead of disease, or dead of other causes. The study

was approved by the Institutional Review Board of

Memorial Sloan-Kettering Cancer Center.

Statistical AnalysisDescriptive statistics were used to summarize study

data. Associations between categorical variables were

evaluated by using the Fisher exact test or the chi-

square test, as appropriate. In all statistical analysis, a

2-tailed P value < .05 was considered statistically sig-

nificant. Outcomes were classified according to sites

of first disease recurrence. Follow-up was calculated

from the time of surgery to date of last follow-up.

RESULTSReclassification of Follicular Lesions and Identificationof Study PatientsThe computerized data base search identified 552

follicular neoplasms with adequate material from

patients who were seen between 1980 and 1995. After

microscopic examination by one of the authors

(R.A.G.), 162 FVPTCs were identified. Of these 162

FVPTCs that were identified by our initial histopatho-

logic examination, only 82 patients were retained as

potential participants for the study after the elimina-

tion of all subcentimeter FVPTCs and FVPTCs that

were associated with >2 foci of papillary microcarci-

noma or other significant thyroid malignancies (see

Material and Methods, above). After rereview by all 4

pathologists, 78 patients finally were included in the

study.

Papillary Thyroid CA Follicular Variant/Liu et al. 1257

Pathologic ParametersTable 1 lists the clinical and pathologic features of the

78 patient who were included in the current study. A

median of 12 slides per patient were reviewed. The

median tumor size was 2.5 cm, and 9 of 78 FVPTCs

(12%) exceeded 4 cm. The neoplasms were encapsu-

lated in 61 of 78 patients (78%), whereas 17 of 78

neoplasms (22%) were nonencapsulated and invaded

the surrounding neoplastic tissue in an infiltrative or

diffuse pattern (Figs. 1, 2). Table 2 lists the clinical

and pathologic features according to the invasive

growth patterns of the FVPTCs (encapsulated vs. non-

encapsulated). There was no significant difference in

tumor size, mitotic rate, presence of a predominantly

oncocytic cytoplasm, or vascular invasion between

patients with encapsulated FVPTCs and patients with

nonencapsulated FVPTCs. In contrast, patients with

encapsulated FVPTCs had a significantly lower rate of

marked intratumoral fibrosis (18% vs. 88% in none-

ncapsulated FVPTCs), extrathyroid extension (5% vs.

65%, respectively), and positive margins (2% vs. 50%,

respectively; P < .0001). The nuclear features charac-

teristic of papillary carcinomas were present diffusely

in 53 of 61 encapsulated FVPTCs (87%) and multifocal

in 8 of 61 encapsulated FVPTCs (13%). We did not

assess the distribution of PTC nuclei in the nonencap-

sulated FVPTCs, because the diagnosis of malignancy

usually is obvious in these tumors based on the inva-

sive properties of the tumor. The presence of any

invasion (capsular or vascular) was observed in 18 of

61 encapsulated FVPTCs (31%; 6 tumors showed both

capsular and vascular invasion, 8 tumors showed cap-

sular invasion only, and 4 tumors showed vascular in-

vasion alone). Forty-three of 61 encapsulated FVPTCs

(70%) were totally noninvasive.

Clinical ParametersThe median age for all 78 patients who were inclu-

ded in this study was 43.1 years (range, 6.5–71.9

years). Fifty-nine of 78 patients (76%) were female.

Fifty-four patients underwent a lobectomy, and 21

patients underwent total thyroidectomy. The remain-

ing 3 patients underwent a lobectomy with partial

resection of the contralateral lobe (2 patients) and a

subtotal thyroidectomy (1 patient). Thirteen patients

underwent formal lymph node dissection, including

3 patients who underwent dissection of the central

compartment only; 7 patients who underwent uni-

lateral, modified neck dissection; 1 patient who

underwent unilateral, modified neck dissection that

included the central compartment; and 2 patients

who underwent bilateral, modified neck dissection

that included the central compartment. Radioactive

TABLE 1Clinical and Pathologic Features of the Follicular Variant of PapillaryThyroid Carcinoma (n = 78 Patients)

Variable

No. of

patients (%)

% of

total

Age, y (median, 43.1 years)

>45 34 44

�45 44 56

Gender

Female 59 76

Male 19 24

Tumor size, cm (median, 2.5 cm)

>4 9 12

�4 69 88

Histologic growth patterns (‘‘subvariants’’)

Encapsulated 61 78

Infiltrative 16 21

Diffuse 1 1

Vascular invasion

No 64 82

Focal (<4 foci) 9 12

Extensive (�4 foci) 5 6

Capsular invasion (n ¼ 61 patients)

No 47 77

Focal (<4 foci) 8 13

Extensive (�4 foci) 6 10

Intratumoral fibrosis

No 41 53

Mild 11 14

Marked 26 33

Mitosis

Absent 76 97

Present 2 3

Oncocytic cytoplasm

Present 11 14

Absent 67 86

Extrathyroid extension

Present 14 18

Absent 64 82

Margins

Positive 7 9

Negative 57 73

Unknown 14 18

Type of thyroid surgery

Lobectomy 54 69

More than lobectomy

but less than

total thyroidectomy

3 4

Total thyroidectomy 21 27

RAI therapy

Administered 15 19

Not given 63 81

Lymph node metastases

Present 14 18

Absent 64 82

Distant metastases

Present 0 0

Absent 78 100

RAI indicates radioactive iodine therapy.

1258 CANCER September 15, 2006 / Volume 107 / Number 6

iodine was given to only 15 of 78 patients (19%). Re-

gional lymph node metastases were present in 14 of

78 patients (18%), and distant metastases were pres-

ent in none. Table 2 lists the clinicopathologic fea-

tures according to the invasive growth pattern

(encapsulated vs. nonencapsulated). The lymph node

metastatic rate was significantly higher in patients

who had nonencapsulated tumors (11 of 17 patients;

65%) compared with patients who had encapsulated

neoplasms (3 of 61 patients; 5%; P < .0001). All 43

patients who had noninvasive, encapsulated FVPTC

lacked evidence of lymph node metastases. The

lymph node metastatic rate was significantly higher

in the infiltrative/diffuse group (65%) than in the

invasive encapsulated group (3 of 18 patients; 17%;

P ¼ .006).Thirty-two patients had lymph node tissue

available for microscopic examination. In those

32 patients, the metastatic lymph node rate was sig-

nificantly higher for patients who had nonencapsu-

lated FVPTC (12 of 13 patients; 92%) than for patients

who had encapsulated FVPTC (3 of 19 patients; 16%;

P < .0001). Among the patients who underwent lym-

phadenectomy, the median age in the encapsulated

group (45.6 years) was very similar to that in the

nonencapsulated group (43 years; P ¼ .72). A similar

correlation was found for tumor size, with a similar

median size of 2.5 cm for both the encapsulated

group and the nonencapsulated group (P ¼ .5). In

patients who had assessable lymph node tissue, there

was a greater proportion of males in the nonencap-

sulated group (6 of 13 patients; 46%) than in the

encapsulated group (3 of 19 patients; 16%), but this

difference was not significant (P ¼ .11). Within the

group with invasive, encapsulated FVPTCs, the lymph

node metastatic rate was significantly higher (3 of 7

patients; 42%) among patients who had �4 foci of

invasion (capsular and/or vascular) than among

patients who had less invasive foci (0 of 11 patients;

P ¼ .04.

Outcome DataAll patients who had >1 year of follow-up were

included in the outcome analysis. Sixty-nine patients

fulfilled this criteria (Table 3) with a median follow-

up of 10.8 years (range, 1.2–21 years). All 42 patients

who had noninvasive, encapsulated FVPTC were free

of disease at follow-up with no recurrence noted and

no lymph node metastases. A female patient age

FIGURE 1. This sample of the noninvasive, encapsulated, follicular variant of papillary thyroid carcinoma (FVPTC), which measured 3.2 cm, was from a malepatient age 51 years who was treated by lobectomy only. The patient did not develop lymph node or distant metastasis and did not develop recurrent disease

after 19 years of follow-up. (A) Low-power view showing the tumor (T), capsule (arrow), and adjacent nonneoplastic tissue (N). (B) Medium-power view showing

dense, pink-staining colloid that often is seen in FVPTC (arrow). (C) High-power view showing clear, overlapping nuclei (arrow). (D) High-power view showing

nuclear grooves (arrow).

Papillary Thyroid CA Follicular Variant/Liu et al. 1259

15 years who had encapsulated FVPTC that showed

capsular invasion (>4 foci) and 2 foci of vascular

invasion developed a recurrence in her neck lymph

nodes 2 years after she underwent total thyroidec-

tomy and received RAI therapy. Thirty-one patients

with noninvasive, encapsulated FVPTC (with >1 year

of follow-up) underwent lobectomy alone and did

not receive RAI therapy. None of those 31 patients

with noninvasive, encapsulated FVPTC developed a

recurrence: This group had a median follow-up of

11.1 years (range, 1.2–21 years), a median tumor size

of 2.3 cm (range, 1.1–7 cm), a median age of 43.4

years (range, 11.5–69.1 years), and a male:female ra-

tio of 1:5.2 (5 men and 26 women) (Table 4). Those

31 patients were subdivided according to the Ameri-

can Joint Committee on Cancer tumor, lymph node,

metastasis (TNM) staging system (6th edition) as fol-

lows: Twenty-one patients had Stage I disease (any T,

any N, M0, and younger than age 45 years or T1 [tu-

mor measuring �2 cm], N0, M0, and age 45 years

or older), 8 patients had Stage II disease (T2 [tumor

measuring >2 cm but <4 cm], N0, M0, and age

45 years or older), and 2 patients had Stage III dis-

ease (T3 [tumor measuring >4 cm], N0, M0, and age

45 years or older).

DISCUSSIONIn the current study, we adopted very stringent crite-

ria for the diagnosis of FVPTC. Indeed, we excluded

many papillary carcinomas from the FVPTC category

if they contained a minor but substantial amount of

papillae. The presence of >1% papillary formations

eliminated the tumor from the FVPTC category. We

also did not include tumors that displayed a high mi-

totic rate and tumor necrosis. This may explain why

we found only 1 diffuse follicular variant in our se-

ries. To understand the biologic behavior of FVPTC,

we excluded all subcentimeter FVPTCs and allowed

only 2 additional foci of papillary microcarcinoma as

additional malignancy. In view of this study design,

we cannot assess multicentricity in FVPTC. There

was no significant disagreement between the pathol-

ogists on the study, possibly because of their com-

mon training at the same institution. The median

age (43.1 years) and the female predominance (76%)

of the patient population were in keeping with previ-

ous studies on papillary carcinomas that included

FVPTCs.2,15 The encapsulated FVPTCs outnumbered

their infiltrative/diffuse counterparts (only 17 tumors

were diffuse or infiltrative vs. 61 encapsulated FVPTCs).

This rarity of infiltrative/diffuse (i.e., nonencapsu-

lated) FVPTC seems to be concordant with the first

detailed article on FVPTCs by Chem and Rosai.16 In

their 1977 article, all their cases were infiltrative with

an apparently very low incidence since they found

only 6 cases of infiltrative FVPTC out of all thyroid

carcinoma cases diagnosed at the University of

Minnesota Hospitals up to the year 1975. Using a less

FIGURE 2. This sample of the infiltra-tive variant of papillary thyroid carcinoma

(FVPTC), which measured 2.3 cm, was

from a male patient age 53 years who

presented with metastatic carcinoma in

13 of 19 cervical lymph nodes. (A) Low-

power view with marked intratumoral fi-

brosis (F). (B) Low-power view of the tu-

mor (T) infiltrating nonneoplastic thyroid

(N). The advancing front of the infiltrating

carcinoma is marked with an arrow. (C)

This view shows the dense, pink-staining

colloid that often is seen in FVPTC

(arrow). (D) High-power view showing

clear nuclei with grooves (arrow).

1260 CANCER September 15, 2006 / Volume 107 / Number 6

stringent definition of FVPTC than Chem and Rosai

by allowing up to 20% papillary formations, Tielens

et al. observed a 5.6% incidence of infiltrative FVPTC

among all well differentiated thyroid carcinomas.4

In the current study, patients who had infiltra-

tive/diffuse FVPTC had a significantly greater fre-

quency (P < .0001) of marked intratumoral fibrosis,

extrathyroid extension, and positive margins than

patients who had encapsulated FVPTC. It is note-

worthy that there was no significant difference in

vascular invasion between encapsulated FVPTCs and

infiltrative/diffuse FVPTCs. This superior potential of

nonencapsulated FVPTCs in invading the thyroid

and extrathyroid stroma was reflected by its higher

rate of total thyroidectomy and especially by its rate

of regional lymph node metastases. Indeed, patients

TABLE 2Clinical and Pathologic Characteristics According to the HistologicSubvariants of Follicular Variant of Papillary Thyroid Carcinoma(Encapsulated, Nonencapsulated)

Characteristic

No. of patients (%)

P*

EncapsulatedFVPTC

(n = 61 patients)

NonencapsulatedFVPTC

(n = 17 patients)

Age, y(median, 41.8 years) .78

�45 33 (54) 10 (59)

>45 28 (46) 7 (41)

Gender

Female 49 (80) 10 (59) .11

Male 12 (20) 7 (41)

Tumor size, cm .19

Median 2.5 2.0

�4 52 (85) 17 (100)

>4 9 (15) 0 (0)

Vascular invasion .49

Absent 51 (84) 13 (76)

Present 10 (16) 4 (24)

Capsular invasion NA

Absent 47 (77) NA

Present 14 (23) NA

PTC nuclei in tumor NA

Multifocal 8 (13) NA

Diffuse 53 (87) NA

Mitosis 1

Absent 59 (97) 17 (100)

Present 2 (3) 0 (0)

Oncocytic cytoplasm .11

Absent 50 (82) 17 (100)

Present 11 (18) 0 (0)

Intratumoral fibrosis <.0001

Absent/mild 50 (82) 2 (12)

Marked 11 (18) 15 (88)

Extrathyroid extension <.0001

Absent 58 (95) 6 (35)

Present 3 (5) 11 (65)

Margins <.0001

Positive 1 (2) 6 (50)

Negative 51 (98) 6 (50)

Thyroid surgery .06

Less than total thyroidectomy 48 (79) 9 (53)

Total thyroidectomy 13 (21) 8 (47)

Lymph node metastases <.0001

Present 3 (5) 11 (65)

Absent 58 (95) 6 (35)

Distant metastases 1

Present 0 (0) 0 (0)

Absent 61 (100) 17 (100)

FVPTC indicates follicular variant of papillary thyroid carcinoma; PTC, papillary thyroid carcinoma;

NA, not applicable.

* Fisher exact test, 2-tailed values. Only patients who had assessable margins were included in the

analysis.

TABLE 3Outcomes According to Histologic Subvariant of Follicular Variant ofPapillary Thyroid Carcinoma (Encapsulated or Nonencapsulated) in69 Patients with Adequate Follow-Up*

Histologic subvariant No. of patients

REC/AWD/DOD:

No of patients (%)

Encapsulated (n ¼ 55 patients)

Noninvasive 42 0 (0)

Invasive 13 1 REC (8)

Nonencapsulated (diffuse/infiltrative) 14 0 (0)

REC indicates recurrence; AWD, alive with disease; DOD, dead of disease.

* Patients were considered to have adequate follow-up if they had >1 year of follow-up. The median

follow-up was 10.8 years.

TABLE 4Outcome and Clinicopathologic Characteristics of 31 Patients withthe Noninvasive, Encapsulated Follicular Variant of Papillary ThyroidCarcinoma with Adequate Follow Up who Underwent Lobectomy Onlywith No Radioactive Iodine Therapy

Characteristic Value

Median follow-up 11.1 y

Median tumor size 2.3 cm

Median age 43.4 y

Male:female ratio 1:5.2

Lymph node metastases Absent

AJCC stage (no. of patients)*

Stage I 21

Stage II 8

Stage III 2

Outcome No recurrence

AJCC indicates American Joint Committee on Cancer (6th edition).

* Patients were classified with Stage I disease if they had either any T, any N, M0, and age <45 years,

or T1 (tumor measuring �2 cm), N0, M0, and age �45 years; they were classified with Stage II dis-

ease if they had T2 (tumor measuring >2 cm but <4 cm), N0, M0, and age �45 years; and they were

classified with Stage III disease if they had T3 (tumor measuring >4 cm), N0, M0, and age �45 years.

Papillary Thyroid CA Follicular Variant/Liu et al. 1261

with nonencapsulated (infiltrative/diffuse) FVPTCs

had a metastatic lymph node rate of 65% compared

with 5% for patients with encapsulated FVPTCs (P <

.0001). This strong and significant correlation between

lymph node metastases and nonencapsulated FVPTC

was maintained when only specimens (32 tumors) that

contained lymph node tissue were analyzed. This dif-

ference in lymph node disease could not be explained

by differences in tumor size or age at presentation,

because the latter 2 variables were similar between

the patients in the encapsulated group and the none-

ncapsulated group. There was a greater proportion of

males in the infiltrative/diffuse group, but this differ-

ence did not reach statistical significance (P ¼ .11).

In the current study, the metastatic lymph node pat-

tern of encapsulated FVPTCs (5%) was much closer

to that reported in follicular carcinomas (on the order

of 5–10%), whereas infiltrative/diffuse FVPTCs had a

metastatic lymph node pattern within the range re-

ported for classic papillary carcinomas (on the order

of 45–65%).11,4,2 Our overall lymph node rate in

FVPTCs (18%) was slightly higher than that reported

by Zhu et al. (13%) but lower than the frequency of

lymph node disease reported by Zidan et al. and Tie-

lens et al. (22%).4,11,17 The latter difference may be

explained by the observation that those investigators

allowed up to 20% of papillae in their FVPTCs. Their

tumors will be now classified as classic PTC by most

authors, because the modern and now universally

accepted description of FVPTC indicates that the tu-

mor must be entirely or almost completely follicular

in pattern.16,18 Among the invasive, encapsulated

FVPTCs, there was a much higher lymph node meta-

static rate in tumors that had �4 foci of invasion

than in tumors that had fewer invasive foci (42% vs.

0%; P ¼ .04). Indeed, the very few patients who had

encapsulated FVPTCs (only 3 patients) that metasta-

sized to lymph nodes all had capsular and vascular

invasion with >10 foci of invasion in each patient. A

similar correlation between lymph node metastases

and extent of invasion was reported in patients who

had follicular carcinomas of the oncocytic subtype.19

With regard to prognosis, patients who had inva-

sive tumors, whether encapsulated or not, had a rare

but real potential for adverse outcome. One patient

who had an encapsulated FVPTC with capsular and

vascular invasion developed a recurrence in the cervi-

cal lymph nodes 2 years after surgery. With a median

follow-up of almost 11 years, overall, 6% of patients

who had invasive tumors, whether encapsulated or

not, had adverse outcomes, whereas none of the 42

patients who had noninvasive, encapsulated FVPTCs

developed recurrences, developed metastases, or died

of disease. All 31 patients who had noninvasive, encap-

sulated FVPTCs and who had adequate follow-up, a

median tumor size of 2.3 cm, and underwent lobec-

tomy alone had good outcomes and no lymph node

metastases (median follow-up, 11.1 years). These data

confirm the view that patients with noninvasive, en-

capsulated FVPTC have an excellent prognosis.6 Eight

of the noninvasive, encapsulated FVPTCs had multi-

focal distribution of the nuclear features of papillary

carcinoma. We classified the entire tumor as FVPTC

according to the recommendation of Baloch and

LiVolsi with the understanding that some pathologists

may count the foci with atypical nuclei and report the

lesion as multifocal PTC.10 Other may use alternative

terminology, such as tumor as of uncertain malignant

potential.20 Whatever position the investigator takes

regarding the nuclear features of FVPTC, the current

results seem to point to the importance of invasion

rather than nuclear features for predicting outcomes

in patients who have encapsulated FVPTC. Indeed,

encapsulated FVPTC seem to have a behavior much

closer to that of follicular tumors (i.e., FTA and FTC)

rather than classic PTC. This is reflected by the lack of

adverse outcomes in patients with noninvasive lesions

and the rarity of lymph node metastases. These mor-

phologic and clinical data are supported by several

publications pointing to a molecular profile of FVPTC

much closer to the FTA/FTC group than to classic

PTC. Zhu et al. reported a low frequency of RET/PTC

rearrangement (3%) and a high frequency of ras muta-

tions in FVPTC (43%) very similar to follicular carci-

noma.11 BRAF mutations reportedly were absent in

FVPTCs and FTCs but present in 53% of classic PTCs.21

The genome-wide profile of FVPTCs assessed by com-

parative genomic hybridization was identified as very

different from classic PTC and closer to the FTA/FTC

group.13 It is believed that PAX8-PPARg rearrange-

ments are restricted to follicular carcinomas and are

absent in papillary carcinomas.22 Most recently, Castro

et al. reported a similar frequency of PAX8-PPARg rear-

rangements in FVPTCs (37.5%), FTCs (45.5%), and

FTAs (33.3%).14 Based on their data and on other stu-

dies,10 those authors concluded that FVPTC as a whole

is not a subgroup of conventional PTC.14 It is note-

worthy that some of those molecular studies were

composed entirely or in their majority of encapsulated

FVPTCs.

The clinical, histologic, and molecular data gath-

ered to date strongly suggest that FVPTC is a hetero-

geneous disease composed of 2 distinct group of

tumors: a nonencapsulated (infiltrative and diffuse)

subvariant, which resembles classic PTC in its inva-

sive growth and metastatic lymph node pattern, and

encapsulated FVPTC. The latter seems to resemble

FTA/FTC in its invasive property, metastatic lymph

1262 CANCER September 15, 2006 / Volume 107 / Number 6

node pattern, and molecular profile. The lack of

adverse outcomes among our 31 patients with non-

invasive, encapsulated FVPTC who underwent lobec-

tomy only (with a median follow-up of 11.1 years

and a median tumor size of 2.3 cm) strongly suggests

that capsular and vascular invasion, and not nuclear

features, is the determinant of malignant behavior in

encapsulated FVPTC. It is interesting to note that

90% (28 of 31 patients) of our patients with noninva-

sive, encapsulated FVPTC would have been treated

by many experts today with total thyroidectomy and

RAI therapy, because these tumors measured �1.5 cm

in size.9 The total lack of recurrence in our group of

patients with noninvasive FVPTC who underwent lo-

bectomy alone suggests that noninvasive, encapsu-

lated FVPTC may be managed by lobectomy only, as

recommended by Rosai et al. in the last Armed Forces

Institute of Pathology fascicle on thyroid tumors.6

Larger studies with longer follow-up than the current

series will be needed to refine therapy for patients

with noninvasive, encapsulated FVPTC. If the current

findings are confirmed, then strong consideration

should be given to reclassifying encapsulated FVPTC

as an entity that is close to the FTA/FTC class of

tumors (Fig. 3). The same criteria that were used to

decide whether follicular tumors are biologically

benign or malignant (i.e., capsular and vascular inva-

sion) would be applied to the evaluation of encapsu-

lated FVPTCs. In practical terms, a lack of capsular

or vascular invasion should denote a benign clinical

behavior in encapsulated FVPTC. If this reclassi-

fication is realized, then it will have a major impact

on the diagnosis and management of patients with

FVPTC. In noninvasive, encapsulated FVPTC, patho-

logists will be spared the frustrating and subjective

exercise of deciding whether a tumor has the nuclear

features of papillary carcinoma. More important,

countless numbers of patients with noninvasive,

encapsulated FVPTC will be spared unnecessary and

aggressive therapy with its attached morbidity (i.e.,

hypoparathyroidism and recurrent nerve injury) and

financial costs.

REFERENCES1. Chan JKC. Tumors of the thyroid and parathyroid glands.

In: Fletcher CDM, editor. Histopathology of Tumors, 2nd

ed. London: Churchill Livingstone; 2000:959–1038.

2. Passler C, Prager G, Scheuba C, et al. Follicular variant of

papillary thyroid carcinoma: a long-term follow-up. Arch

Surg. 2003;138:1362–1366.

3. Carcangiu ML, Zampi G, Pupi A, et al. Papillary carcinoma

of the thyroid. A clinicopathologic study of 241 cases trea-

ted at the University of Florence, Italy. Cancer. 1985;55:805–

828.

4. Tielens ET, Sherman SI, Hruban RH, et al. Follicular variant

of papillary thyroid carcinoma. A clinicopathologic study.

Cancer. 1994;73:424–431.

5. Sebastian SO, Gonzalez JM, Paricio PP, et al. Papillary thyr-

oid carcinoma: prognostic index for survival including the

histological variety. Arch Surg. 2000;135:272–277.

6. Rosai J, Carcangiu ML, Delellis RA. Follicular carcinomas.

Papillary Carcinomas. In: Rosai J, Sobin LH. Tumors of the

thyroid gland. Atlas of Tumor Pathology. Washington, DC:

Armed Forces Institute of Pathology, 1992:49–121.

7. Chan JK. Strict criteria should be applied in the diagnosis

of encapsulated follicular variant of papillary thyroid carci-

noma. Am J Clin Pathol. 2002;117:16–18.

8. Lloyd RV, Erickson LA, Casey MB, et al. Observer variation

in the diagnosis of follicular variant of papillary thyroid

carcinoma. Am J Surg Pathol. 2004;28:1336–1340.

9. Solomon BL, Wartofsky L, Burman KD. Current trends in

the management of well differentiated papillary thyroid

carcinoma. J Clin Endocrinol Metab. 1996;81:333–339.

10. Baloch ZW, LiVolsi VA. Encapsulated follicular variant of

papillary thyroid carcinoma with bone metastases. Mod

Pathol. 2000;13:861–865.

11. Zhu Z, Gandhi M, Nikiforova MN, et al. Molecular profile

and clinical-pathologic features of the follicular variant of

papillary thyroid carcinoma. An unusually high prevalence

of ras mutations. Am J Clin Pathol. 2003;120:71–77.

12. Jain M, Khan A, Patwardhan N, et al. Follicular variant of

papillary thyroid carcinoma: a comparative study of histo-

pathologic features and cytology results in 141 patients.

Endocr Pract. 2001;7:79–84.

13. Wreesmann VB, Ghossein RA, Hezel M, et al. Follicular vari-

ant of papillary thyroid carcinoma: genome-wide appraisal

of a controversial entity. Genes Chromosomes Cancer. 2004;40:

355–364.

14. Castro P, Rebocho AP, Soares RJ, et al. PAX8-PPARgamma

rearrangement is frequently detected in the follicular vari-

ant of papillary thyroid carcinoma. J Clin Endocrinol

Metab. 2006;91:213–220.

15. Rosai J, Zampi G, Carcangiu ML. Papillary carcinoma of the

thyroid. A discussion of its several morphologic expressions,

FIGURE 3. The suggested relationship between the follicular variant ofpapillary thyroid carcinoma (FVPTC), classic papillary thyroid carcinoma

(PTC), and follicular thyroid adenoma/follicular thyroid carcinoma (FTA/FTC) is

illustrated based on the histologic, clinical, and molecular data available to

date. Nonencapsulated (infiltrative/diffuse) PTC is very close to classic PTC,

whereas encapsulated FVPTC is close to FTA/FTC.

Papillary Thyroid CA Follicular Variant/Liu et al. 1263

with particular emphasis on the follicular variant. Am J Surg

Pathol. 1983;7:809–817.

16. Chem KT, Rosai J. Follicular variant of thyroid papillary

carcinoma: a clinicopathologic study of six cases. Am J

Surg Pathol. 1977;1:123–130.

17. Zidan J, Karen D, Stein M, et al. Pure versus follicular vari-

ant of papillary thyroid carcinoma: clinical features, pro-

gnostic factors, treatment, and survival. Cancer. 2003;97:

1181–1185.

18. LiVolsi V. Pure versus follicular variant of papillary thyroid

carcinoma: clinical features, prognostic factors [comment].

Cancer. 2003;98:1997.

19. Stojadinovic A, Ghossein RA, Hoos A, et al. Hurthle cell car-

cinoma: a critical histopathologic appraisal. J Clin Oncol. 2001;

19:2616–2625.

20. Williams ED. Guest editorial: two proposals regarding the ter-

minology of thyroid tumors. Int J Surg Pathol. 2000;8:181–183.

21. Trovisco V, Vieira de Castro I, Soares P, et al. BRAF muta-

tions are associated with some histological types of papil-

lary thyroid carcinoma. J Pathol. 2004;202:247–251.

22. Nikiforova MN, Biddinger PW, Caudill CM, et al. PAX8-

PPARgamma rearrangement in thyroid tumors: RT-PCR

and immunohistochemical analyses. Am J Surg Pathol. 2002;

26:1016–1023.

1264 CANCER September 15, 2006 / Volume 107 / Number 6