Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Original Research n Ultrasonography
Malignancy Risk Stratification in Thyroid Nodules with Nondiagnostic Results at Cytologic Examination: Combination of Thyroid Imaging Reporting and Data System and the Bethesda System1 Hee Jung Moon, MD, PhD Eun-Kyung Kim, MD, PhD Jung Hyun Yoon, MD Jin Young Kwak, MD, PhD
1
From the Department of Radiology, Research Institute of Radiological Science, Severance Hospital, Yonsei University College of Medicine, 50 Yonse-ro, Seodaemun-gu, 120-752 Seoul, Korea. Received February 11, 2014; revision requested March 21; revision received May 7; accepted May 19; final version accepted June 11. Address correspondence to J.Y.K. (e-mail: [email protected]).
Purpose:
To evaluate the malignancy risks of thyroid nodules with nondiagnostic results at ultrasonography (US)–guided fine-needle aspiration biopsy (FNAB) and the criteria for selecting those for repeat US-guided FNAB according to the thyroid imaging reporting and data system (TIRADS).
Materials and Methods:
This retrospective study was approved by the institutional review board, and the requirement to obtain informed consent was waived. Five hundred forty-eight nondiagnostic nodules were included. US features of internal composition, echogenicity, margin, calcifications, shape, and vascularity were evaluated, and thyroid nodules were classified according to TIRADS. TIRADS category 3 included nodules without any suspicious features of solidity, hypoechogenicity or marked hypoechogenicity, microlobulated or irregular margins, microcalcifications, and tallerthan-wide shape. Categories 4a, 4b, 4c, and 5 included nodules with one, two, three or four, or five suspicious US features. The malignancy risk was calculated.
Results:
Of the 548 nodules, 40 (7.3%) were malignant and 508 (92.7%) were benign. The malignancy risks of categories 3 and 4a nodules were 0.8% and 1.8%, respectively, whereas the malignancy risks of categories 4b, 4c, and 5 nodules were 6.1%, 14.4%, and 31%. In the 294 nodules larger than 10 mm, the malignancy risks of categories 3, 4a, 4b, 4c, and 5 nodules were 0.9%, 1.3%, 0%, 15%, and 33%, respectively. In the 254 nodules measuring 10 mm or smaller, the malignancy risks of categories 3, 4a 4b, 4c, and 5 nodules were 0%, 2.7%, 14%, 14.3%, and 31%.
Conclusion:
Nondiagnostic thyroid nodules without suspicious US features and those with one suspicious feature can be followed up with US, but nondiagnostic nodules with two or more suspicious features should undergo repeat USguided FNAB. q RSNA, 2014
RSNA, 2014
q
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
287
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
F
ine-needle aspiration biopsy (FNAB) is a standard method for triaging thryoid nodules to surgery or clinical follow-up, and, with FNAB, the number of unnecessary surgeries has decreased (1,2). The main limitation of FNAB, however, is nondiagnostic or unsatisfactory results. According to the Bethesda System for Reporting Thyroid Cytopathology (3), a sample is considered nondiagnostic or unsatisfactory when the specimen shows obscuring blood, overlying thick smears, air drying of alcohol-fixed smears, or an inadequate number of follicular cells. According to the Bethesda system (3), nondiagnostic results should ideally be limited to less than 10% of all thyroid FNABs, but the rates of nondiagnostic results are reported to be as high as
Advances in Knowledge nn The malignancy risks of nondiagnostic thyroid nodules classified as thyroid imaging reporting and data system (TIRADS) categories 3 and 4a were 0.8% (one of 130 nodules) and 1.8% (two of 113 nodules), respectively, whereas the risks for categories 4b, 4c, and 5 nodules were 6.1% (seven of 115 nodules), 14.4% (25 of 174 nodules), and 31% (five of 16 nodules). nn In the 294 nodules larger than 10 mm, the malignancy risks for categories 3 and 4a nodules were 0.9% (one of 110 nodules) and 1.3% (one of 76 nodules), whereas the malignancy risks for categories 4b, 4c, and 5 nodules were 0% (0 of 64 nodules), 15% (six of 41 nodules), and 33% (one of three nodules). nn In the 254 nodules measuring 10 mm or smaller, the malignancy risks for categories 3 and 4a nodules were 0% (0 of 20 nodules) and 2.7% (one of 37 nodules), whereas the malignancy risks for categories 4b, 4c, and 5 nodules were 14% (seven of 51 nodules), 14.3% (19 of 133 nodules), and 31% (four of 13 nodules). 288
Moon et al
21% (4–10). Some authors have demonstrated that the best option for reducing nondiagnostic results is on-site cytologic assessment (11,12). However, even with on-site assessment by cytopathologists, the prevalence of nondiagnostic results has still been reported to be 10.7% (13). The recommended risk of malignancy for thyroid nodules with nondiagnostic cytologic findings with the Bethesda system is 1%–4%, where the denominator is the total number of nondiagnostic results and the numerator is the number of pathologically proved malignancies (2,3,10,14). If we change the denominator into cases with repeat FNAB and surgery and the numerator into cases with proved malignancies, the malignancy risk increases from 14% to 51% (1,2,6,10,14,15). In these cases, a substantial portion of thyroid nodules with nondiagnostic results did not undergo surgery or repeat FNAB, with selection bias being thought the cause for the increased and wide range of malignancy risk. In most studies, excluding one by Baloch et al (6), who reported a 51% risk of malignancy, 81%–84% of nodules with nondiagnostic results that underwent repeat FNAB or surgery were finally proved to be benign (1,2,4,7,10). If nondiagnostic nodules without pathologic or clinical evidence of malignancy are considered benign with use of the same calculation method as in the Bethesda system (3), malignancy rates for nondiagnostic nodules in previous reports are calculated to range from 0.6% to 10.5%, with a pooled average of 90.5% of nodules benign in these reports (1,2,6,10,14). Therefore, for more effective management and accurate prediction of malignancy, risk stratification in nondiagnostic thyroid nodules is needed.
Implication for Patient Care nn Nondiagnostic thyroid nodules without suspicious US features or those with one suspicious US feature can be followed up with US; nondiagnostic nodules with two or more suspicious features should undergo repeat US-guided fine-needle aspiration biopsy.
The thyroid imaging reporting and data system (TIRADS) developed by Kwak et al (16) can help stratify thyroid nodules according to malignancy risk by using the number of suspicious ultrasonography (US) features such as solidity, hypoechogenicity or marked hypoechogenicity, microlobulated or irregular margins, microcalcifications, and tallerthan-wide shape. TIRADS can help accurately predict malignancy and can be easily applied in clinical practice owing to its simplicity (16,17). However, this reporting system has not been applied to thyroid nodules with nondiagnostic results at cytologic examination, even when it can be used in the continuous risk stratification of nodules in this category after FNAB. Therefore, we evaluated the malignancy risks of thyroid nodules with nondiagnostic results at US-guided FNAB biopsy and the criteria for selecting those for repeat USguided FNAB according to the TIRADS.
Materials and Methods Study Population Our institutional review board approved this retrospective study, and the requirement to obtain informed consent was waived. From January 2010 to November 2010, 4767 thyroid nodules measuring at least 5 mm underwent US-guided FNAB. Eight hundred thirty-three of the Published online before print 10.1148/radiol.14140359 Content codes: Radiology 2015; 274:287–295 Abbreviations: FNAB = fine-needle aspiration biopsy TIRADS = thyroid imaging reporting and data system Author contributions: Guarantors of integrity of entire study, H.J.M., J.Y.K.; study concepts/study design or data acquisition or data analysis/ interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, H.J.M.; clinical studies, H.J.M., E.K.K., J.Y.K.; statistical analysis, H.J.M.; and manuscript editing, H.J.M., E.K.K., J.H.Y. Conflicts of interest are listed at the end of this article.
radiology.rsna.org n Radiology: Volume 274: Number 1—January 2015
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
Moon et al
Figure 1
Figure 1: Diagram of study group. a = 39 of 79 nodules had a cystic portion, b = six of 14 nodules had a cystic portion.
4767 nodules (17.5%) had nondiagnostic cytologic findings according to the Bethesda system (3). Of the 833 nodules, 756 (90.7%) underwent initial USguided FNAB during the study period and had nondiagnostic results. Of the 756 nodules, those fulfilling the following conditions were included: nodules with diagnostic results at repeat US-guided FNAB (n = 185), those with nondiagnostic results at repeat US-guided FNAB showing no change during at least 12 months of follow-up US at our institution (mean follow-up, 33.8 months; range, 12–50.6 months) (n = 47), those with nondiagnostic results at repeat USguided FNAB showing decreased size at follow-up US at our institution (n = 14), nodules showing no change during at least 12 months of follow-up US at our institution (mean follow-up, 31.6 months; range, 12–47.1 months) (n = 155), nodules showing decreased size at follow-up US at our institution (n = 79), and nodules that had undergone surgery (n = 68). Of the 93 nodules that decreased in size, 45 had a cystic portion within the nodules in which the decrease in size might have been due to the aspiration of cyst contents. We also used
data from the National Cancer Center Registry to classify benign and malignant nodules. Two hundred eight nodules that did not undergo repeat FNAB, surgery, or at least 12 months of clinical and US follow-up were excluded. In all, a total of 548 nodules in 530 patients were included in this study (Fig 1). Of the 530 patients, 430 were women and 100 were men. The mean patient age (6standard deviation) was 52.7 years 6 11.8 (median age, 53 years; range, 17–81 years). The mean age of female patients was 52.3 years (range, 17–81 years), and the mean age of male patients was 54.1 years (range, 19–78 years) (P = .178). The mean size of the 548 thyroid nodules was 16 mm 6 11.7 (median size, 12 mm; range, 5–60 mm). The mean time between initial FNAB and repeat FNAB was 10.3 months 6 9.2 (median, 6.2 months; range, 0.3–47.6 months). The mean interval between the initial FNAB and the last follow-up US examination at our institution was 31.6 months 6 10.7 (median, 34.5 months; range, 12–47.1 months). The mean interval between initial FNAB and surgery was 9.1 months 6 11.3 (median, 4.5 months; range, 0.4–46.5 months).
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
US Examination and US-guided FNAB US was performed by using high-spatial-resolution US machines equipped with either a 5–12-MHz linear probe (iU22; Philips Medical Systems, Bothell, Wash) or a 6–14-MHz linear probe (EUB-7500; Hitachi Medical, Tokyo, Japan). US examinations were performed by one of nine radiologists with 1–16 years of experience in thyroid imaging. All US-guided FNABs were performed by the same radiologist who performed the US examinations. Nodule size was defined by using the maximum diameter at US. US features of thyroid nodules that underwent US-guided FNAB were prospectively recorded according to internal composition, echogenicity, margin, calcifications, shape, and vascularity at the time of US-guided FNAB (16,18,19). The internal composition was classified as solid, cystic portion of 50% or less, and cystic portion greater than 50%. The nodule was classified as hyper-, iso-, or hypoechogenic compared with normal thyroid gland or as showing marked hypoechogenicity when a nodule was relatively hypoechoic compared with the surrounding strap muscle. Echogenicity 289
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
was classified according to the predominant pattern if the nodule was heterogeneous and the solid portion if the nodule had cystic portions. Margins were classified as well defined, microlobulated, or irregular. Calcifications, if present, were classified as microcalcifications (1 mm in diameter; tiny, punctate, hyperechoic foci with or without acoustic shadows) or macrocalcifications. Nodules with both microcalcifications and macrocalcifications were classified as microcalcifications. Shape was classified as wider than tall or taller than wide (greater in its anteroposterior dimension than in its transverse dimension). Vascularity was classified as peripheral (flow visualized with power Doppler US only at the periphery of the nodule and not in the tissue peripheral to the nodule), central (flow visualized with power Doppler US within the nodule regardless of the presence of flow at the periphery of the nodule), and no vascularity (no power Doppler flow in the periphery and within the nodule) (19). Suspicious malignant US features included marked hypoechogenicity, microlobulated or irregular margin, microcalcifications, and taller-than-wide shape (18,19). USguided FNAB was performed in nodules with suspicious features or in the largest thyroid nodule without any suspicious features. In our institution, US-guided FNAB was performed in symptomatic simple cysts only, and these were not included in this study. US-guided FNAB was performed with the capillary fill and free-hand technique without an aspirator by using a 23-gauge needle attached to a 2-mL disposable plastic syringe. Many passages were performed until aspirated materials were seen in the needle hub. Samples obtained in this matter were expelled onto two different glass slides, which were immediately placed in 95% alcohol for Papanicolaou staining. Afterward, with use of a new 23-gauge needle and 2-mL syringe, aspiration was repeated to make another two slides. Samples were smeared after each aspiration, and the remaining aspirated material was rinsed with saline and processed for cell blocking. Pathologists studied cell blocks before reaching a final diagnosis. Cytopathologists were 290
Moon et al
Table 1 Comparison of Demographic Characteristics and US Features in All Thyroid Nodules according to Malignancy and Benignity Parameter Mean age (y)* Sex Female Male Mean size (mm)* Composition Solid Cystic portion 50% Cystic portion .50% Echogenicity Hyperechogenicity Isoechogenicity Hypoechogenicity Marked hypoechogenicity Margin Well defined Microlobulated Irregular Calcifications No calcifications Microcalcifications Macrocalcifications Shape Wider than tall Taller than wide Vascularity Peripheral Central None TIRADS category 3 4a 4b 4c 5
All Nodules (n = 548)
Malignant Nodules (n = 40)
Benign Nodules (n = 508)
...
48.2 (9.8)
53 (11.9)
446 102 ...
37 (8.3) 3 (2.9) 9.7 (7.1)
409 (91.7) 99 (97.1) 16.3 (11.8)
352 126 70
39 (11.1) 1 (0.8) 0 (0)
313 (88.9) 125 (99.2) 70 (100)
2 205 319 22
0 (0) 4 (2) 27 (8.5) 9 (41)
2 (100) 201 (98) 292 (91.5) 13 (59)
374 95 79
15 (4) 10 (10) 15 (19)
359 (96) 85 (89) 64 (81)
401 64 83
19 (4.7) 13 (20) 8 (9.6)
382 (95.3) 51 (80) 75 (90)
459 89
24 (5.2) 16 (18)
435 (94.8) 73 (82)
186 163 199
13 (7) 6 (3.7) 21 (10.6)
173 (93) 157 (96.3) 178 (89.4)
130 113 115 174 16
1 (0.8) 2 (1.8) 7 (6.1) 25 (14.4) 5 (31)
129 (99.2) 111 (98.2) 108 (93.9) 149 (85.6) 11 (69)
P Value .013 .096
,.0001 ,.0001
,.0001
,.0001
,.0001
,.0001
.037
,.0001
Note.—Except where indicated, numbers in parentheses are percentages. * Numbers in parentheses are standard deviations.
not on-site during biopsies. The cytologic results were reported according to the Bethesda classification, which has been used in our institution since December 2009 (3).
Data and Statistical Analysis Thyroid nodules with malignant results at surgery were classified as malignant. Thyroid nodules with benign results at surgery, without malignant cytologic
findings at repeat US-guided FNAB, with no change during at least 12 months of follow-up US after nondiagnostic results at US-guided FNAB, or decreased size at follow-up were classified as benign. The 295 cases without surgery or diagnostic results at repeat US-guided FNAB were classified as benign after considering data collected from clinical follow-ups at our institution and from the National Cancer Center Registry. In
radiology.rsna.org n Radiology: Volume 274: Number 1—January 2015
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
TIRADS, solidity, hypoechogenicity or marked hypoechogenicity, microlobulated or irregular margins, microcalcifications, and taller-than-wide shape were defined as suspicious malignant US features (16). Thyroid nodules without suspicious features were classified as TIRADS category 3 (16). Thyroid nodules with one, two, three or four, or five suspicious features were classified as category 4a, 4b, 4c, or 5, respectively (16). Thyroid nodules included in this study were reclassified according to TIRADS. The malignancy risk of each TIRADS category was calculated and presented as a percentage. Age and nodule size were compared between benign and malignant nodules by using the independent t test or Mann Whitney U test. Sex, US features, and TIRADS categories were compared between benign and malignant nodules by using the x2 or Fisher exact test. Statistical analyses with the methods mentioned earlier were done for all thyroid nodules, for nodules larger than 10 mm, and for nodules measuring 5–10 mm. Statistical analysis was performed by using software (version 20.0; SPSS, Chicago, Ill). Two-sided P , .05 was considered indicative of a statistically significant difference.
Results Of the 548 nodules, 40 (7.3%) were malignant and 508 (92.7%) were benign. When the denominator included all of the 756 initially nondiagnostic nodules, the malignancy risk was 5.3% (40 of 756 nodules). Surgery was performed for all 40 malignancies. The mean age of patients with malignancy was 48.2 years, and the mean age of patients with benign lesions was 53 years (P = .013) (Table 1). The distribution of sexes was not significantly different between patients with malignant and benign nodules (P = .096). The US feature of solid composition was more frequently seen in malignant nodules (P , .0001). The malignancy risk of nodules with a cystic portion greater than 50% was 0% (0 of 70 nodules). Of 70 nodules with a cystic portion greater than 50%, 43 were
Moon et al
Figure 2
Figure 2: Nondiagnostic thyroid nodule with cystic portion greater than 50%. Transverse US scan in 53-year-old woman with 8-mm thyroid nodule in right thyroid gland. Nodule has cystic portion greater than 50%, hypoechogenicity, microlobulated margin, and taller-than-wide shape. With its three suspicious features, the nodule was assessed as TIRADS category 4c. Patient underwent surgery due to thyroid malignancy in her left thyroid and the nodule was ultimately diagnosed as benign.
assessed as TIRADS category 3, 21 as category 4a, three as category 4b, and three as category 4c (Fig 2). Marked hypoechogenicity, microlobulated or irregular margins, microcalcifications, and taller-than-wide shape were more frequently seen in malignant nodules (P , .0001 for all) (Fig 3). No vascularity was more frequently seen in malignant nodules (P = .037). The malignancy risks of categories 3, 4a, 4b, 4c, and 5 nodules were 0.8% (one of 130 nodules), 1.8% (two of 113 nodules), 6.1%, (seven of 115 nodules), 14.4% (25 of 174 nodules), and 31% (five of 16 nodules), respectively (P , .0001) (Fig 4). Of the 294 nodules larger than 10 mm, nine (3.1%) were malignant and 285 (96.9%) were benign. Solid composition was substantially seen in malignant nodules (P = .069). Marked hypoechogenicity, microlobulated or irregular margin, and microcalcifications were more frequently seen in malignant nodules (P = .010, .005, and .001, respectively) (Table 2). Shape and vascularity did not show significant differences between malignant and benign nodules (P = .109 and .139, respectively). The malignancy risks of categories 3, 4a, 4b, 4c, and 5 nodules
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
were 0.9% (one of 110 nodules), 1.3% (one of 76 nodules), 0% (0 of 64 nodules), 15% (six of 41 nodules), and 33% (one of three nodules), respectively (P , .0001) (Fig 4). Of the 254 nodules measuring 5–10 mm, 31 (12.2%) were malignant and 223 (87.8%) were benign. Solid composition was more frequently seen in malignant nodules (P = .019). Marked hypoechogenicity and taller-than-wide shape were more frequently seen in malignant nodules (P , .0001 and P = .033, respectively) (Table 3). Margin, calcifications, and vascularity did not show significant differences between malignant and benign nodules (P = .052, .135, and .99, respectively). The malignancy risks of categories 3, 4a, 4b, 4c, and 5 nodules were 0% (0 of 20 nodules), 2.7% (one of 37 nodules), 14% (seven of 51 nodules), 14.3% (19 of 133 nodules), and 31% (four of 13 nodules) (P = .023) (Fig 4).
Discussion The malignancy risk of nondiagnostic thyroid nodules in our study was 5.3% (40 of 756 nodules) when the denominator was all nondiagnostic nodules 291
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
Moon et al
Figure 3
Figure 3: Initial nondiagnostic solid thyroid nodule shown to be malignant at repeat FNAB. US scan in 41-year-old woman with 14-mm thyroid nodule in left thyroid gland. Nodule is solid and has hypoechogenicity, microlobulated margin, micro- and macrocalcifications, and taller-than-wide shape. With its five suspicious features, the nodule was assessed as TIRADS category 5. Patient underwent repeat FNAB after 3 months, and malignancy was diagnosed at cytologic and histopathologic examinations.
Figure 4
Figure 4: Bar chart shows malignancy risks of all nondiagnostic thyroid nodules according to TIRADS. cat = category.
and 7.3% (40 of 548 nodules) when the denominator was nondiagnostic nodules with surgery, repeat US-guided FNAB, or at least 12 months of follow-up US. These risk values are within 292
the 0.6%–10.5% risk range published in previous studies (2,6,10,14,15) but are slightly higher than the 1%–4% recommended in the Bethesda system (3). When the 294 nondiagnostic nodules
larger than 10 mm were analyzed separately, 3.1% (nine of 294 nodules) were malignant; this is within the 1%– 4% malignancy risk range given in the Bethesda system (3). However, the malignancy risk of the 254 nondiagnostic nodules measuring 10 mm or less was 12.2% (31 of 254 nodules), which was much higher than that recommended in the Bethesda system (3) as well as the 3.1% risk found in nodules larger than 10 mm. This discrepancy arose because US-guided FNAB was performed more selectively on suspicious thyroid nodules within the 5–10-mm range than on those larger than 10 mm. None of the nondiagnostic nodules with a cystic portion larger than 50% at US was proved malignant among the 548 nondiagnostic nodules included in our study. In a study by Renshaw (10), cysts found at cytologic examination were analyzed separately from nondiagnostic results. The malignancy risk was 3.9% (three of 77 nodules) when the denominator included all nondiagnostic nodules and 13.8% when the denominator included only nodules with surgery or repeat FNAB (10). In that study, however, two-thirds of FNABs were performed with palpation guidance and the remaining one-third under imaging guidance (10). Thus, the proportion of cystic composition within the nodules could not be analyzed and was not described (10). In other studies about cystic thyroid nodules based on US features, the malignancy risks of nodules with a cystic portion greater than 50% at US were reported as 2.2% (20) and 17.6% (21). In the Bethesda system, if a nondiagnostic thyroid nodule is entirely cystic and shows no suspicious US features, the thyroid nodule is regarded as benign and can be recommended for clinical follow-up (3). As the solid portion within the nodules increases, the malignancy risk also increases (20). In our study, thyroid nodules were classified according to the cystic portion within the nodule; none of the nodules with a cystic portion larger than 50% were malignant regardless of nodule size and other US features. Thus, nondiagnostic nodules with a cystic portion larger than 50% might be considered for US follow-up.
radiology.rsna.org n Radiology: Volume 274: Number 1—January 2015
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
Moon et al
Table 2 Comparison of Demographic Characteristics and US Features in Thyroid Nodules Larger than 10 mm according to Malignancy and Benignity Parameter Composition Solid Cystic portion 50% Cystic portion .50% Echogenicity Hyperechogenicity Isoechogenicity Hypoechogenicity Marked hypoechogenicity Margin Well defined Microlobulated Irregular Calcifications No calcifications Microcalcifications Macrocalcifications Shape Wider than tall Taller than wide Vascularity Peripheral Central None TIRADS category 3 4a 4b 4c 5
All Nodules (n = 294)
Malignant Nodules (n = 9)
Benign Nodules (n = 285)
141 98 55
8 (5.7) 1 (1) 0 (0)
133 (94.3) 97 (99) 55 (100)
0 158 127 9
0 (0) 2 (1.3) 5 (3.9) 2 (22)
0 (0) 156 (98.7) 122 (96.1) 7 (78)
252 23 19
4 (1.6) 3 (13) 2 (10)
248 (98.4) 20 (87) 17 (89)
234 15 45
3 (1.3) 3 (20) 3 (6.7)
231 (98.7) 12 (80) 42 (93)
275 19
7 (2.5) 2 (10)
268 (97.5) 17 (89)
96 125 73
2 (2.1) 2 (1.6) 5 (6.8)
94 (98) 123 (98.4) 68 (93)
110 76 64 41 3
1 (0.9) 1 (1.3) 0 (0) 6 (15) 1 (33)
109 (99.1) 75 (99) 64 (100) 35 (85) 2 (67)
P Value .069
.010
.005
.001
.109
.139
,.0001
Note.—Numbers in parentheses are percentages.
Neither of the two hyperechoic nodules in our study was malignant, but this is too small of a sample to make conclusions regarding this feature. In nondiagnostic thyroid nodules with isoechogenicity and those with a cystic portion of 50% or less, the malignancy risks were 2% (four of 205 nodules) and 0.8% (one of 126 nodules), respectively. These risks were within the 0%–3% malignancy risk range in benign nodules that would indicate clinical follow-up (3). Other US features, except cystic portion of 50% or less, cystic portion greater than 50%, and isoechogenicity, showed malignancy risks higher than 3%.
The TIRADS developed by Kwak et al (16) enables stratification of thyroid nodules according to malignancy risk. TIRADS category 3 in the study by Kwak et al is given to thyroid nodules without suspicious features, and categories 4a, 4b, 4c, and 5 are given to nodules with one, two, three or four, or five suspicious features, respectively (16). Malignancy risks for each TIRADS category have been reported to be 1.7%, 3.3%, 9.2%, 44.4%–72.4%, and 87.5%, respectively (16). In all 548 nondiagnostic nodules, the malignancy risks of nondiagnostic nodules classified as category 3 and category 4a were 0.8% (one of 130 nodules)
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
and 1.8% (two of 113 nodules), respectively. The 0.8% and 1.8% were within the 0%–3% range, which is the risk range for recommending clinical follow-up in benign nodules with the Bethesda system (3). Thus, US follow-up can be sufficient for nondiagnostic nodules if the nodules have no suspicious US features or one suspicious US feature. Conversely, the malignancy risks of categories 4b, 4c, and 5 thyroid nodules were 6.1% (seven of 115 nodules), 14.4% (25 of 174 nodules), and 31.3% (five of 16 nodules), respectively. Thus, repeat US-guided FNAB should be performed for those nodules. In the 294 nodules larger than 10 mm, the malignancy risks of nondiagnostic nodules classified as category 3 and 4a were 0.9% (one of 110 nodules) and 1.3% (one of 76 nodules), whereas the malignancy risks of category 4c and 5 nodules were 15% (six of 41 nodules) and 33% (one of three nodules), respectively. In the 254 nodules measuring at least 10 mm, the malignancy risks of nondiagnostic nodules classified as category 3 and 4a were 0% (0 of 20 nodules) and 2.7% (one of 37 nodules), but the malignancy risks of those classified as category 4b, 4c, and 5 were 14% (seven of 51 nodules), 14.3% (19 of 133 nodules), and 31% (four of 13 nodules), respectively. Repeat US-guided FNAB should be performed for nondiagnostic thyroid nodules classified as category 4b, 4c, and 5 regardless of nodule size, but nodules classified as category 3 and 4a can be followed up with US. The optimal time for performing repeat US-guided FNAB after nondiagnostic results was not evaluated in our study. In the Bethesda system, the optimal time for repeat FNAB is not described clearly, although FNAB is recommended after 3 months owing to reparative atypia; some authors recommend FNAB within the next 6–12 months (3,6,12). The optimal time for US follow-up for nondiagnostic nodules with a cystic portion larger than 50% or those classified as category 3 or 4a was also not investigated. There are some limitations to our study. First, our study was of a 293
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
Moon et al
Disclosures of Conflicts of Interest: H.J.M. disclosed no relevant relationships. E.K.K. disclosed no relevant relationships. J.H.Y. disclosed no relevant relationships. J.Y.K. disclosed no relevant relationships.
Table 3 Comparison of Demographic Characteristics and US Features in Thyroid Nodules 10 mm or Smaller according to Malignancy and Benignity Parameter Composition Solid Cystic portion 50% Cystic portion .50% Echogenicity Hyperechogenicity Isoechogenicity Hypoechogenicity Marked hypoechogenicity Margin Well defined Microlobulated Irregular Calcifications No calcifications Microcalcifications Macrocalcifications Shape Wider than tall Taller than wide Vascularity Peripheral Central None TIRADS category 3 4a 4b 4c 5
All Nodules (n = 254)
Malignant Nodules (n = 31)
Benign Nodules (n = 223)
P Value
References
.019 211 28 15
31 (14.7) 0 (0) 0 (0)
180 (85.3) 28 (100) 15 (100)
2 47 192 13
0 (0) 2 (4.3) 22 (11.5) 7 (54)
2 (100) 45 (96) 170 (88.5) 6 (46)
122 72 60
11 (9) 7 (9.7) 13 (22)
111 (91) 65 (90) 47 (78)
167 38 49
16 (9.6) 5 (13.2) 10 (20)
151 (90.4) 33 (87) 39 (79)
184 70
17 (9.2) 14 (20)
167 (90.8) 56 (80)
1. Richards ML, Bohnenblust E, Sirinek K, Bingener J. Nondiagnostic thyroid fineneedle aspiration biopsies are no longer a dilemma. Am J Surg 2008;196(3):398–402.
,.0001
2. Yang J, Schnadig V, Logrono R, Wasserman PG. Fine-needle aspiration of thyroid nodules: a study of 4703 patients with histologic and clinical correlations. Cancer 2007;111(5):306–315.
.052
3. Cibas ES, Ali SZ; NCI Thyroid FNA State of the Science Conference. The Bethesda system for reporting thyroid cytopathology. Am J Clin Pathol 2009;132(5):658–665.
.135
4. Alexander EK, Heering JP, Benson CB, et al. Assessment of nondiagnostic ultrasound-guided fine needle aspirations of thyroid nodules. J Clin Endocrinol Metab 2002;87(11):4924–4927.
.033
5. Baier ND, Hahn PF, Gervais DA, et al. Fineneedle aspiration biopsy of thyroid nodules: experience in a cohort of 944 patients. AJR Am J Roentgenol 2009;193(4):1175–1179.
.99 90 38 126
11 (12) 4 (10) 16 (12.7)
79 (88) 34 (89) 110 (87.3)
6. Baloch Z, LiVolsi VA, Jain P, et al. Role of repeat fine-needle aspiration biopsy (FNAB) in the management of thyroid nodules. Diagn Cytopathol 2003;29(4):203–206.
.023 20 37 51 133 13
0 (0) 1 (2.7) 7 (14) 19 (14.3) 4 (31)
20 (100) 36 (97) 44 (86) 114 (85.7) 9 (69)
7. Chow LS, Gharib H, Goellner JR, van Heerden JA. Nondiagnostic thyroid fineneedle aspiration cytology: management dilemmas. Thyroid 2001;11(12): 1147–1151.
Note.—Numbers in parentheses are percentages.
retrospective design, and 208 of 756 nondiagnostic nodules (27.5%) were excluded because the US follow-up interval was less than 12 months. In addition, the malignancy risk of the 254 nondiagnostic nodules measuring 10 mm or less was 12.2% (31 of 254 nodules) because US-guided FNAB was performed more selectively on suspicious thyroid nodules within the 5–10-mm range than in those larger than 10 mm. Thus, a selection bias was present. Second, we included thyroid nodules with follow-up US or clinical follow-up that had not undergone repeat US-guided FNAB or surgery. Thyroid nodules without cytologic, histologic, or clinical evidence 294
of malignancy were considered benign. Third, 548 nodules in 530 patients were included. All analysis was performed per nodule and not per patient, which could have resulted in overestimation of our findings. However, a generalized estimating equation analysis was not considered because there were data cells with zero value for composition, echogenicity, and TIRADS categories. In conclusion, nondiagnostic thyroid nodules without suspicious US features or those with one suspicious US feature can be followed up with US, but nondiagnostic nodules with two or more suspicious features should undergo repeat US-guided FNAB.
8. Marqusee E, Benson CB, Frates MC, et al. Usefulness of ultrasonography in the management of nodular thyroid disease. Ann Intern Med 2000;133(9):696–700. 9. Moon HJ, Kwak JY, Kim EK, Kim MJ. Ultrasonographic characteristics predictive of nondiagnostic results for fine-needle aspiration biopsies of thyroid nodules. Ultrasound Med Biol 2011;37(4):549–555. 10. Renshaw AA. Accuracy of thyroid fineneedle aspiration using receiver operator characteristic curves. Am J Clin Pathol 2001;116(4):477–482. 11. Redman R, Zalaznick H, Mazzaferri EL, Massoll NA. The impact of assessing specimen adequacy and number of needle passes for fine-needle aspiration biopsy of thyroid nodules. Thyroid 2006;16(1): 55–60.
radiology.rsna.org n Radiology: Volume 274: Number 1—January 2015
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
12. Jaume JC, Chen H. Inadequate cytology of thyroid nodules: repeat it or live with it. Indian J Surg Oncol 2011;2(2):76–77. 13. Jing X, Wey E, Michael CW. Retrospective evaluation of instituted standard adequacy criteria for on-site adequacy assessment of thyroid fine-needle aspiration. Diagn Cytopathol 2011;39(6):391–394. 14. Ravetto C, Colombo L, Dottorini ME. Usefulness of fine-needle aspiration in the diagnosis of thyroid carcinoma: a retrospective study in 37,895 patients. Cancer 2000;90(6):357–363. 15. Yoon JH, Moon HJ, Kim EK, Kwak JY. Inadequate cytology in thyroid nodules: should
Moon et al
we repeat aspiration or follow-up? Ann Surg Oncol 2011;18(5):1282–1289.
ble solid nodules of the thyroid. AJR Am J Roentgenol 2002;178(3):687–691.
16. Kwak JY, Han KH, Yoon JH, et al. Thyroid imaging reporting and data system for US features of nodules: a step in establishing better stratification of cancer risk. Radiology 2011;260(3):892–899.
19. Moon HJ, Kwak JY, Kim MJ, Son EJ, Kim EK. Can vascularity at power Doppler US help predict thyroid malignancy? Radiology 2010;255(1):260–269.
17. Kwak JY, Jung I, Baek JH, et al. Image reporting and characterization system for ultrasound features of thyroid nodules: multicentric Korean retrospective study. Korean J Radiol 2013;14(1):110–117. 18. Kim EK, Park CS, Chung WY, et al. New sonographic criteria for recommending fine-needle aspiration biopsy of nonpalpa-
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
20. Lee MJ, Kim EK, Kwak JY, Kim MJ. Partially cystic thyroid nodules on ultrasound: probability of malignancy and sonographic differentiation. Thyroid 2009;19(4):341–346. 21. Bellantone R, Lombardi CP, Raffaelli M, et al. Management of cystic or predominantly cystic thyroid nodules: the role of ultrasound-guided fine-needle aspiration biopsy. Thyroid 2004;14(1):43–47.
295
Original Research n Ultrasonography
Malignancy Risk Stratification in Thyroid Nodules with Nondiagnostic Results at Cytologic Examination: Combination of Thyroid Imaging Reporting and Data System and the Bethesda System1 Hee Jung Moon, MD, PhD Eun-Kyung Kim, MD, PhD Jung Hyun Yoon, MD Jin Young Kwak, MD, PhD
1
From the Department of Radiology, Research Institute of Radiological Science, Severance Hospital, Yonsei University College of Medicine, 50 Yonse-ro, Seodaemun-gu, 120-752 Seoul, Korea. Received February 11, 2014; revision requested March 21; revision received May 7; accepted May 19; final version accepted June 11. Address correspondence to J.Y.K. (e-mail: [email protected]).
Purpose:
To evaluate the malignancy risks of thyroid nodules with nondiagnostic results at ultrasonography (US)–guided fine-needle aspiration biopsy (FNAB) and the criteria for selecting those for repeat US-guided FNAB according to the thyroid imaging reporting and data system (TIRADS).
Materials and Methods:
This retrospective study was approved by the institutional review board, and the requirement to obtain informed consent was waived. Five hundred forty-eight nondiagnostic nodules were included. US features of internal composition, echogenicity, margin, calcifications, shape, and vascularity were evaluated, and thyroid nodules were classified according to TIRADS. TIRADS category 3 included nodules without any suspicious features of solidity, hypoechogenicity or marked hypoechogenicity, microlobulated or irregular margins, microcalcifications, and tallerthan-wide shape. Categories 4a, 4b, 4c, and 5 included nodules with one, two, three or four, or five suspicious US features. The malignancy risk was calculated.
Results:
Of the 548 nodules, 40 (7.3%) were malignant and 508 (92.7%) were benign. The malignancy risks of categories 3 and 4a nodules were 0.8% and 1.8%, respectively, whereas the malignancy risks of categories 4b, 4c, and 5 nodules were 6.1%, 14.4%, and 31%. In the 294 nodules larger than 10 mm, the malignancy risks of categories 3, 4a, 4b, 4c, and 5 nodules were 0.9%, 1.3%, 0%, 15%, and 33%, respectively. In the 254 nodules measuring 10 mm or smaller, the malignancy risks of categories 3, 4a 4b, 4c, and 5 nodules were 0%, 2.7%, 14%, 14.3%, and 31%.
Conclusion:
Nondiagnostic thyroid nodules without suspicious US features and those with one suspicious feature can be followed up with US, but nondiagnostic nodules with two or more suspicious features should undergo repeat USguided FNAB. q RSNA, 2014
RSNA, 2014
q
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
287
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
F
ine-needle aspiration biopsy (FNAB) is a standard method for triaging thryoid nodules to surgery or clinical follow-up, and, with FNAB, the number of unnecessary surgeries has decreased (1,2). The main limitation of FNAB, however, is nondiagnostic or unsatisfactory results. According to the Bethesda System for Reporting Thyroid Cytopathology (3), a sample is considered nondiagnostic or unsatisfactory when the specimen shows obscuring blood, overlying thick smears, air drying of alcohol-fixed smears, or an inadequate number of follicular cells. According to the Bethesda system (3), nondiagnostic results should ideally be limited to less than 10% of all thyroid FNABs, but the rates of nondiagnostic results are reported to be as high as
Advances in Knowledge nn The malignancy risks of nondiagnostic thyroid nodules classified as thyroid imaging reporting and data system (TIRADS) categories 3 and 4a were 0.8% (one of 130 nodules) and 1.8% (two of 113 nodules), respectively, whereas the risks for categories 4b, 4c, and 5 nodules were 6.1% (seven of 115 nodules), 14.4% (25 of 174 nodules), and 31% (five of 16 nodules). nn In the 294 nodules larger than 10 mm, the malignancy risks for categories 3 and 4a nodules were 0.9% (one of 110 nodules) and 1.3% (one of 76 nodules), whereas the malignancy risks for categories 4b, 4c, and 5 nodules were 0% (0 of 64 nodules), 15% (six of 41 nodules), and 33% (one of three nodules). nn In the 254 nodules measuring 10 mm or smaller, the malignancy risks for categories 3 and 4a nodules were 0% (0 of 20 nodules) and 2.7% (one of 37 nodules), whereas the malignancy risks for categories 4b, 4c, and 5 nodules were 14% (seven of 51 nodules), 14.3% (19 of 133 nodules), and 31% (four of 13 nodules). 288
Moon et al
21% (4–10). Some authors have demonstrated that the best option for reducing nondiagnostic results is on-site cytologic assessment (11,12). However, even with on-site assessment by cytopathologists, the prevalence of nondiagnostic results has still been reported to be 10.7% (13). The recommended risk of malignancy for thyroid nodules with nondiagnostic cytologic findings with the Bethesda system is 1%–4%, where the denominator is the total number of nondiagnostic results and the numerator is the number of pathologically proved malignancies (2,3,10,14). If we change the denominator into cases with repeat FNAB and surgery and the numerator into cases with proved malignancies, the malignancy risk increases from 14% to 51% (1,2,6,10,14,15). In these cases, a substantial portion of thyroid nodules with nondiagnostic results did not undergo surgery or repeat FNAB, with selection bias being thought the cause for the increased and wide range of malignancy risk. In most studies, excluding one by Baloch et al (6), who reported a 51% risk of malignancy, 81%–84% of nodules with nondiagnostic results that underwent repeat FNAB or surgery were finally proved to be benign (1,2,4,7,10). If nondiagnostic nodules without pathologic or clinical evidence of malignancy are considered benign with use of the same calculation method as in the Bethesda system (3), malignancy rates for nondiagnostic nodules in previous reports are calculated to range from 0.6% to 10.5%, with a pooled average of 90.5% of nodules benign in these reports (1,2,6,10,14). Therefore, for more effective management and accurate prediction of malignancy, risk stratification in nondiagnostic thyroid nodules is needed.
Implication for Patient Care nn Nondiagnostic thyroid nodules without suspicious US features or those with one suspicious US feature can be followed up with US; nondiagnostic nodules with two or more suspicious features should undergo repeat US-guided fine-needle aspiration biopsy.
The thyroid imaging reporting and data system (TIRADS) developed by Kwak et al (16) can help stratify thyroid nodules according to malignancy risk by using the number of suspicious ultrasonography (US) features such as solidity, hypoechogenicity or marked hypoechogenicity, microlobulated or irregular margins, microcalcifications, and tallerthan-wide shape. TIRADS can help accurately predict malignancy and can be easily applied in clinical practice owing to its simplicity (16,17). However, this reporting system has not been applied to thyroid nodules with nondiagnostic results at cytologic examination, even when it can be used in the continuous risk stratification of nodules in this category after FNAB. Therefore, we evaluated the malignancy risks of thyroid nodules with nondiagnostic results at US-guided FNAB biopsy and the criteria for selecting those for repeat USguided FNAB according to the TIRADS.
Materials and Methods Study Population Our institutional review board approved this retrospective study, and the requirement to obtain informed consent was waived. From January 2010 to November 2010, 4767 thyroid nodules measuring at least 5 mm underwent US-guided FNAB. Eight hundred thirty-three of the Published online before print 10.1148/radiol.14140359 Content codes: Radiology 2015; 274:287–295 Abbreviations: FNAB = fine-needle aspiration biopsy TIRADS = thyroid imaging reporting and data system Author contributions: Guarantors of integrity of entire study, H.J.M., J.Y.K.; study concepts/study design or data acquisition or data analysis/ interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, H.J.M.; clinical studies, H.J.M., E.K.K., J.Y.K.; statistical analysis, H.J.M.; and manuscript editing, H.J.M., E.K.K., J.H.Y. Conflicts of interest are listed at the end of this article.
radiology.rsna.org n Radiology: Volume 274: Number 1—January 2015
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
Moon et al
Figure 1
Figure 1: Diagram of study group. a = 39 of 79 nodules had a cystic portion, b = six of 14 nodules had a cystic portion.
4767 nodules (17.5%) had nondiagnostic cytologic findings according to the Bethesda system (3). Of the 833 nodules, 756 (90.7%) underwent initial USguided FNAB during the study period and had nondiagnostic results. Of the 756 nodules, those fulfilling the following conditions were included: nodules with diagnostic results at repeat US-guided FNAB (n = 185), those with nondiagnostic results at repeat US-guided FNAB showing no change during at least 12 months of follow-up US at our institution (mean follow-up, 33.8 months; range, 12–50.6 months) (n = 47), those with nondiagnostic results at repeat USguided FNAB showing decreased size at follow-up US at our institution (n = 14), nodules showing no change during at least 12 months of follow-up US at our institution (mean follow-up, 31.6 months; range, 12–47.1 months) (n = 155), nodules showing decreased size at follow-up US at our institution (n = 79), and nodules that had undergone surgery (n = 68). Of the 93 nodules that decreased in size, 45 had a cystic portion within the nodules in which the decrease in size might have been due to the aspiration of cyst contents. We also used
data from the National Cancer Center Registry to classify benign and malignant nodules. Two hundred eight nodules that did not undergo repeat FNAB, surgery, or at least 12 months of clinical and US follow-up were excluded. In all, a total of 548 nodules in 530 patients were included in this study (Fig 1). Of the 530 patients, 430 were women and 100 were men. The mean patient age (6standard deviation) was 52.7 years 6 11.8 (median age, 53 years; range, 17–81 years). The mean age of female patients was 52.3 years (range, 17–81 years), and the mean age of male patients was 54.1 years (range, 19–78 years) (P = .178). The mean size of the 548 thyroid nodules was 16 mm 6 11.7 (median size, 12 mm; range, 5–60 mm). The mean time between initial FNAB and repeat FNAB was 10.3 months 6 9.2 (median, 6.2 months; range, 0.3–47.6 months). The mean interval between the initial FNAB and the last follow-up US examination at our institution was 31.6 months 6 10.7 (median, 34.5 months; range, 12–47.1 months). The mean interval between initial FNAB and surgery was 9.1 months 6 11.3 (median, 4.5 months; range, 0.4–46.5 months).
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
US Examination and US-guided FNAB US was performed by using high-spatial-resolution US machines equipped with either a 5–12-MHz linear probe (iU22; Philips Medical Systems, Bothell, Wash) or a 6–14-MHz linear probe (EUB-7500; Hitachi Medical, Tokyo, Japan). US examinations were performed by one of nine radiologists with 1–16 years of experience in thyroid imaging. All US-guided FNABs were performed by the same radiologist who performed the US examinations. Nodule size was defined by using the maximum diameter at US. US features of thyroid nodules that underwent US-guided FNAB were prospectively recorded according to internal composition, echogenicity, margin, calcifications, shape, and vascularity at the time of US-guided FNAB (16,18,19). The internal composition was classified as solid, cystic portion of 50% or less, and cystic portion greater than 50%. The nodule was classified as hyper-, iso-, or hypoechogenic compared with normal thyroid gland or as showing marked hypoechogenicity when a nodule was relatively hypoechoic compared with the surrounding strap muscle. Echogenicity 289
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
was classified according to the predominant pattern if the nodule was heterogeneous and the solid portion if the nodule had cystic portions. Margins were classified as well defined, microlobulated, or irregular. Calcifications, if present, were classified as microcalcifications (1 mm in diameter; tiny, punctate, hyperechoic foci with or without acoustic shadows) or macrocalcifications. Nodules with both microcalcifications and macrocalcifications were classified as microcalcifications. Shape was classified as wider than tall or taller than wide (greater in its anteroposterior dimension than in its transverse dimension). Vascularity was classified as peripheral (flow visualized with power Doppler US only at the periphery of the nodule and not in the tissue peripheral to the nodule), central (flow visualized with power Doppler US within the nodule regardless of the presence of flow at the periphery of the nodule), and no vascularity (no power Doppler flow in the periphery and within the nodule) (19). Suspicious malignant US features included marked hypoechogenicity, microlobulated or irregular margin, microcalcifications, and taller-than-wide shape (18,19). USguided FNAB was performed in nodules with suspicious features or in the largest thyroid nodule without any suspicious features. In our institution, US-guided FNAB was performed in symptomatic simple cysts only, and these were not included in this study. US-guided FNAB was performed with the capillary fill and free-hand technique without an aspirator by using a 23-gauge needle attached to a 2-mL disposable plastic syringe. Many passages were performed until aspirated materials were seen in the needle hub. Samples obtained in this matter were expelled onto two different glass slides, which were immediately placed in 95% alcohol for Papanicolaou staining. Afterward, with use of a new 23-gauge needle and 2-mL syringe, aspiration was repeated to make another two slides. Samples were smeared after each aspiration, and the remaining aspirated material was rinsed with saline and processed for cell blocking. Pathologists studied cell blocks before reaching a final diagnosis. Cytopathologists were 290
Moon et al
Table 1 Comparison of Demographic Characteristics and US Features in All Thyroid Nodules according to Malignancy and Benignity Parameter Mean age (y)* Sex Female Male Mean size (mm)* Composition Solid Cystic portion 50% Cystic portion .50% Echogenicity Hyperechogenicity Isoechogenicity Hypoechogenicity Marked hypoechogenicity Margin Well defined Microlobulated Irregular Calcifications No calcifications Microcalcifications Macrocalcifications Shape Wider than tall Taller than wide Vascularity Peripheral Central None TIRADS category 3 4a 4b 4c 5
All Nodules (n = 548)
Malignant Nodules (n = 40)
Benign Nodules (n = 508)
...
48.2 (9.8)
53 (11.9)
446 102 ...
37 (8.3) 3 (2.9) 9.7 (7.1)
409 (91.7) 99 (97.1) 16.3 (11.8)
352 126 70
39 (11.1) 1 (0.8) 0 (0)
313 (88.9) 125 (99.2) 70 (100)
2 205 319 22
0 (0) 4 (2) 27 (8.5) 9 (41)
2 (100) 201 (98) 292 (91.5) 13 (59)
374 95 79
15 (4) 10 (10) 15 (19)
359 (96) 85 (89) 64 (81)
401 64 83
19 (4.7) 13 (20) 8 (9.6)
382 (95.3) 51 (80) 75 (90)
459 89
24 (5.2) 16 (18)
435 (94.8) 73 (82)
186 163 199
13 (7) 6 (3.7) 21 (10.6)
173 (93) 157 (96.3) 178 (89.4)
130 113 115 174 16
1 (0.8) 2 (1.8) 7 (6.1) 25 (14.4) 5 (31)
129 (99.2) 111 (98.2) 108 (93.9) 149 (85.6) 11 (69)
P Value .013 .096
,.0001 ,.0001
,.0001
,.0001
,.0001
,.0001
.037
,.0001
Note.—Except where indicated, numbers in parentheses are percentages. * Numbers in parentheses are standard deviations.
not on-site during biopsies. The cytologic results were reported according to the Bethesda classification, which has been used in our institution since December 2009 (3).
Data and Statistical Analysis Thyroid nodules with malignant results at surgery were classified as malignant. Thyroid nodules with benign results at surgery, without malignant cytologic
findings at repeat US-guided FNAB, with no change during at least 12 months of follow-up US after nondiagnostic results at US-guided FNAB, or decreased size at follow-up were classified as benign. The 295 cases without surgery or diagnostic results at repeat US-guided FNAB were classified as benign after considering data collected from clinical follow-ups at our institution and from the National Cancer Center Registry. In
radiology.rsna.org n Radiology: Volume 274: Number 1—January 2015
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
TIRADS, solidity, hypoechogenicity or marked hypoechogenicity, microlobulated or irregular margins, microcalcifications, and taller-than-wide shape were defined as suspicious malignant US features (16). Thyroid nodules without suspicious features were classified as TIRADS category 3 (16). Thyroid nodules with one, two, three or four, or five suspicious features were classified as category 4a, 4b, 4c, or 5, respectively (16). Thyroid nodules included in this study were reclassified according to TIRADS. The malignancy risk of each TIRADS category was calculated and presented as a percentage. Age and nodule size were compared between benign and malignant nodules by using the independent t test or Mann Whitney U test. Sex, US features, and TIRADS categories were compared between benign and malignant nodules by using the x2 or Fisher exact test. Statistical analyses with the methods mentioned earlier were done for all thyroid nodules, for nodules larger than 10 mm, and for nodules measuring 5–10 mm. Statistical analysis was performed by using software (version 20.0; SPSS, Chicago, Ill). Two-sided P , .05 was considered indicative of a statistically significant difference.
Results Of the 548 nodules, 40 (7.3%) were malignant and 508 (92.7%) were benign. When the denominator included all of the 756 initially nondiagnostic nodules, the malignancy risk was 5.3% (40 of 756 nodules). Surgery was performed for all 40 malignancies. The mean age of patients with malignancy was 48.2 years, and the mean age of patients with benign lesions was 53 years (P = .013) (Table 1). The distribution of sexes was not significantly different between patients with malignant and benign nodules (P = .096). The US feature of solid composition was more frequently seen in malignant nodules (P , .0001). The malignancy risk of nodules with a cystic portion greater than 50% was 0% (0 of 70 nodules). Of 70 nodules with a cystic portion greater than 50%, 43 were
Moon et al
Figure 2
Figure 2: Nondiagnostic thyroid nodule with cystic portion greater than 50%. Transverse US scan in 53-year-old woman with 8-mm thyroid nodule in right thyroid gland. Nodule has cystic portion greater than 50%, hypoechogenicity, microlobulated margin, and taller-than-wide shape. With its three suspicious features, the nodule was assessed as TIRADS category 4c. Patient underwent surgery due to thyroid malignancy in her left thyroid and the nodule was ultimately diagnosed as benign.
assessed as TIRADS category 3, 21 as category 4a, three as category 4b, and three as category 4c (Fig 2). Marked hypoechogenicity, microlobulated or irregular margins, microcalcifications, and taller-than-wide shape were more frequently seen in malignant nodules (P , .0001 for all) (Fig 3). No vascularity was more frequently seen in malignant nodules (P = .037). The malignancy risks of categories 3, 4a, 4b, 4c, and 5 nodules were 0.8% (one of 130 nodules), 1.8% (two of 113 nodules), 6.1%, (seven of 115 nodules), 14.4% (25 of 174 nodules), and 31% (five of 16 nodules), respectively (P , .0001) (Fig 4). Of the 294 nodules larger than 10 mm, nine (3.1%) were malignant and 285 (96.9%) were benign. Solid composition was substantially seen in malignant nodules (P = .069). Marked hypoechogenicity, microlobulated or irregular margin, and microcalcifications were more frequently seen in malignant nodules (P = .010, .005, and .001, respectively) (Table 2). Shape and vascularity did not show significant differences between malignant and benign nodules (P = .109 and .139, respectively). The malignancy risks of categories 3, 4a, 4b, 4c, and 5 nodules
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
were 0.9% (one of 110 nodules), 1.3% (one of 76 nodules), 0% (0 of 64 nodules), 15% (six of 41 nodules), and 33% (one of three nodules), respectively (P , .0001) (Fig 4). Of the 254 nodules measuring 5–10 mm, 31 (12.2%) were malignant and 223 (87.8%) were benign. Solid composition was more frequently seen in malignant nodules (P = .019). Marked hypoechogenicity and taller-than-wide shape were more frequently seen in malignant nodules (P , .0001 and P = .033, respectively) (Table 3). Margin, calcifications, and vascularity did not show significant differences between malignant and benign nodules (P = .052, .135, and .99, respectively). The malignancy risks of categories 3, 4a, 4b, 4c, and 5 nodules were 0% (0 of 20 nodules), 2.7% (one of 37 nodules), 14% (seven of 51 nodules), 14.3% (19 of 133 nodules), and 31% (four of 13 nodules) (P = .023) (Fig 4).
Discussion The malignancy risk of nondiagnostic thyroid nodules in our study was 5.3% (40 of 756 nodules) when the denominator was all nondiagnostic nodules 291
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
Moon et al
Figure 3
Figure 3: Initial nondiagnostic solid thyroid nodule shown to be malignant at repeat FNAB. US scan in 41-year-old woman with 14-mm thyroid nodule in left thyroid gland. Nodule is solid and has hypoechogenicity, microlobulated margin, micro- and macrocalcifications, and taller-than-wide shape. With its five suspicious features, the nodule was assessed as TIRADS category 5. Patient underwent repeat FNAB after 3 months, and malignancy was diagnosed at cytologic and histopathologic examinations.
Figure 4
Figure 4: Bar chart shows malignancy risks of all nondiagnostic thyroid nodules according to TIRADS. cat = category.
and 7.3% (40 of 548 nodules) when the denominator was nondiagnostic nodules with surgery, repeat US-guided FNAB, or at least 12 months of follow-up US. These risk values are within 292
the 0.6%–10.5% risk range published in previous studies (2,6,10,14,15) but are slightly higher than the 1%–4% recommended in the Bethesda system (3). When the 294 nondiagnostic nodules
larger than 10 mm were analyzed separately, 3.1% (nine of 294 nodules) were malignant; this is within the 1%– 4% malignancy risk range given in the Bethesda system (3). However, the malignancy risk of the 254 nondiagnostic nodules measuring 10 mm or less was 12.2% (31 of 254 nodules), which was much higher than that recommended in the Bethesda system (3) as well as the 3.1% risk found in nodules larger than 10 mm. This discrepancy arose because US-guided FNAB was performed more selectively on suspicious thyroid nodules within the 5–10-mm range than on those larger than 10 mm. None of the nondiagnostic nodules with a cystic portion larger than 50% at US was proved malignant among the 548 nondiagnostic nodules included in our study. In a study by Renshaw (10), cysts found at cytologic examination were analyzed separately from nondiagnostic results. The malignancy risk was 3.9% (three of 77 nodules) when the denominator included all nondiagnostic nodules and 13.8% when the denominator included only nodules with surgery or repeat FNAB (10). In that study, however, two-thirds of FNABs were performed with palpation guidance and the remaining one-third under imaging guidance (10). Thus, the proportion of cystic composition within the nodules could not be analyzed and was not described (10). In other studies about cystic thyroid nodules based on US features, the malignancy risks of nodules with a cystic portion greater than 50% at US were reported as 2.2% (20) and 17.6% (21). In the Bethesda system, if a nondiagnostic thyroid nodule is entirely cystic and shows no suspicious US features, the thyroid nodule is regarded as benign and can be recommended for clinical follow-up (3). As the solid portion within the nodules increases, the malignancy risk also increases (20). In our study, thyroid nodules were classified according to the cystic portion within the nodule; none of the nodules with a cystic portion larger than 50% were malignant regardless of nodule size and other US features. Thus, nondiagnostic nodules with a cystic portion larger than 50% might be considered for US follow-up.
radiology.rsna.org n Radiology: Volume 274: Number 1—January 2015
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
Moon et al
Table 2 Comparison of Demographic Characteristics and US Features in Thyroid Nodules Larger than 10 mm according to Malignancy and Benignity Parameter Composition Solid Cystic portion 50% Cystic portion .50% Echogenicity Hyperechogenicity Isoechogenicity Hypoechogenicity Marked hypoechogenicity Margin Well defined Microlobulated Irregular Calcifications No calcifications Microcalcifications Macrocalcifications Shape Wider than tall Taller than wide Vascularity Peripheral Central None TIRADS category 3 4a 4b 4c 5
All Nodules (n = 294)
Malignant Nodules (n = 9)
Benign Nodules (n = 285)
141 98 55
8 (5.7) 1 (1) 0 (0)
133 (94.3) 97 (99) 55 (100)
0 158 127 9
0 (0) 2 (1.3) 5 (3.9) 2 (22)
0 (0) 156 (98.7) 122 (96.1) 7 (78)
252 23 19
4 (1.6) 3 (13) 2 (10)
248 (98.4) 20 (87) 17 (89)
234 15 45
3 (1.3) 3 (20) 3 (6.7)
231 (98.7) 12 (80) 42 (93)
275 19
7 (2.5) 2 (10)
268 (97.5) 17 (89)
96 125 73
2 (2.1) 2 (1.6) 5 (6.8)
94 (98) 123 (98.4) 68 (93)
110 76 64 41 3
1 (0.9) 1 (1.3) 0 (0) 6 (15) 1 (33)
109 (99.1) 75 (99) 64 (100) 35 (85) 2 (67)
P Value .069
.010
.005
.001
.109
.139
,.0001
Note.—Numbers in parentheses are percentages.
Neither of the two hyperechoic nodules in our study was malignant, but this is too small of a sample to make conclusions regarding this feature. In nondiagnostic thyroid nodules with isoechogenicity and those with a cystic portion of 50% or less, the malignancy risks were 2% (four of 205 nodules) and 0.8% (one of 126 nodules), respectively. These risks were within the 0%–3% malignancy risk range in benign nodules that would indicate clinical follow-up (3). Other US features, except cystic portion of 50% or less, cystic portion greater than 50%, and isoechogenicity, showed malignancy risks higher than 3%.
The TIRADS developed by Kwak et al (16) enables stratification of thyroid nodules according to malignancy risk. TIRADS category 3 in the study by Kwak et al is given to thyroid nodules without suspicious features, and categories 4a, 4b, 4c, and 5 are given to nodules with one, two, three or four, or five suspicious features, respectively (16). Malignancy risks for each TIRADS category have been reported to be 1.7%, 3.3%, 9.2%, 44.4%–72.4%, and 87.5%, respectively (16). In all 548 nondiagnostic nodules, the malignancy risks of nondiagnostic nodules classified as category 3 and category 4a were 0.8% (one of 130 nodules)
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
and 1.8% (two of 113 nodules), respectively. The 0.8% and 1.8% were within the 0%–3% range, which is the risk range for recommending clinical follow-up in benign nodules with the Bethesda system (3). Thus, US follow-up can be sufficient for nondiagnostic nodules if the nodules have no suspicious US features or one suspicious US feature. Conversely, the malignancy risks of categories 4b, 4c, and 5 thyroid nodules were 6.1% (seven of 115 nodules), 14.4% (25 of 174 nodules), and 31.3% (five of 16 nodules), respectively. Thus, repeat US-guided FNAB should be performed for those nodules. In the 294 nodules larger than 10 mm, the malignancy risks of nondiagnostic nodules classified as category 3 and 4a were 0.9% (one of 110 nodules) and 1.3% (one of 76 nodules), whereas the malignancy risks of category 4c and 5 nodules were 15% (six of 41 nodules) and 33% (one of three nodules), respectively. In the 254 nodules measuring at least 10 mm, the malignancy risks of nondiagnostic nodules classified as category 3 and 4a were 0% (0 of 20 nodules) and 2.7% (one of 37 nodules), but the malignancy risks of those classified as category 4b, 4c, and 5 were 14% (seven of 51 nodules), 14.3% (19 of 133 nodules), and 31% (four of 13 nodules), respectively. Repeat US-guided FNAB should be performed for nondiagnostic thyroid nodules classified as category 4b, 4c, and 5 regardless of nodule size, but nodules classified as category 3 and 4a can be followed up with US. The optimal time for performing repeat US-guided FNAB after nondiagnostic results was not evaluated in our study. In the Bethesda system, the optimal time for repeat FNAB is not described clearly, although FNAB is recommended after 3 months owing to reparative atypia; some authors recommend FNAB within the next 6–12 months (3,6,12). The optimal time for US follow-up for nondiagnostic nodules with a cystic portion larger than 50% or those classified as category 3 or 4a was also not investigated. There are some limitations to our study. First, our study was of a 293
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
Moon et al
Disclosures of Conflicts of Interest: H.J.M. disclosed no relevant relationships. E.K.K. disclosed no relevant relationships. J.H.Y. disclosed no relevant relationships. J.Y.K. disclosed no relevant relationships.
Table 3 Comparison of Demographic Characteristics and US Features in Thyroid Nodules 10 mm or Smaller according to Malignancy and Benignity Parameter Composition Solid Cystic portion 50% Cystic portion .50% Echogenicity Hyperechogenicity Isoechogenicity Hypoechogenicity Marked hypoechogenicity Margin Well defined Microlobulated Irregular Calcifications No calcifications Microcalcifications Macrocalcifications Shape Wider than tall Taller than wide Vascularity Peripheral Central None TIRADS category 3 4a 4b 4c 5
All Nodules (n = 254)
Malignant Nodules (n = 31)
Benign Nodules (n = 223)
P Value
References
.019 211 28 15
31 (14.7) 0 (0) 0 (0)
180 (85.3) 28 (100) 15 (100)
2 47 192 13
0 (0) 2 (4.3) 22 (11.5) 7 (54)
2 (100) 45 (96) 170 (88.5) 6 (46)
122 72 60
11 (9) 7 (9.7) 13 (22)
111 (91) 65 (90) 47 (78)
167 38 49
16 (9.6) 5 (13.2) 10 (20)
151 (90.4) 33 (87) 39 (79)
184 70
17 (9.2) 14 (20)
167 (90.8) 56 (80)
1. Richards ML, Bohnenblust E, Sirinek K, Bingener J. Nondiagnostic thyroid fineneedle aspiration biopsies are no longer a dilemma. Am J Surg 2008;196(3):398–402.
,.0001
2. Yang J, Schnadig V, Logrono R, Wasserman PG. Fine-needle aspiration of thyroid nodules: a study of 4703 patients with histologic and clinical correlations. Cancer 2007;111(5):306–315.
.052
3. Cibas ES, Ali SZ; NCI Thyroid FNA State of the Science Conference. The Bethesda system for reporting thyroid cytopathology. Am J Clin Pathol 2009;132(5):658–665.
.135
4. Alexander EK, Heering JP, Benson CB, et al. Assessment of nondiagnostic ultrasound-guided fine needle aspirations of thyroid nodules. J Clin Endocrinol Metab 2002;87(11):4924–4927.
.033
5. Baier ND, Hahn PF, Gervais DA, et al. Fineneedle aspiration biopsy of thyroid nodules: experience in a cohort of 944 patients. AJR Am J Roentgenol 2009;193(4):1175–1179.
.99 90 38 126
11 (12) 4 (10) 16 (12.7)
79 (88) 34 (89) 110 (87.3)
6. Baloch Z, LiVolsi VA, Jain P, et al. Role of repeat fine-needle aspiration biopsy (FNAB) in the management of thyroid nodules. Diagn Cytopathol 2003;29(4):203–206.
.023 20 37 51 133 13
0 (0) 1 (2.7) 7 (14) 19 (14.3) 4 (31)
20 (100) 36 (97) 44 (86) 114 (85.7) 9 (69)
7. Chow LS, Gharib H, Goellner JR, van Heerden JA. Nondiagnostic thyroid fineneedle aspiration cytology: management dilemmas. Thyroid 2001;11(12): 1147–1151.
Note.—Numbers in parentheses are percentages.
retrospective design, and 208 of 756 nondiagnostic nodules (27.5%) were excluded because the US follow-up interval was less than 12 months. In addition, the malignancy risk of the 254 nondiagnostic nodules measuring 10 mm or less was 12.2% (31 of 254 nodules) because US-guided FNAB was performed more selectively on suspicious thyroid nodules within the 5–10-mm range than in those larger than 10 mm. Thus, a selection bias was present. Second, we included thyroid nodules with follow-up US or clinical follow-up that had not undergone repeat US-guided FNAB or surgery. Thyroid nodules without cytologic, histologic, or clinical evidence 294
of malignancy were considered benign. Third, 548 nodules in 530 patients were included. All analysis was performed per nodule and not per patient, which could have resulted in overestimation of our findings. However, a generalized estimating equation analysis was not considered because there were data cells with zero value for composition, echogenicity, and TIRADS categories. In conclusion, nondiagnostic thyroid nodules without suspicious US features or those with one suspicious US feature can be followed up with US, but nondiagnostic nodules with two or more suspicious features should undergo repeat US-guided FNAB.
8. Marqusee E, Benson CB, Frates MC, et al. Usefulness of ultrasonography in the management of nodular thyroid disease. Ann Intern Med 2000;133(9):696–700. 9. Moon HJ, Kwak JY, Kim EK, Kim MJ. Ultrasonographic characteristics predictive of nondiagnostic results for fine-needle aspiration biopsies of thyroid nodules. Ultrasound Med Biol 2011;37(4):549–555. 10. Renshaw AA. Accuracy of thyroid fineneedle aspiration using receiver operator characteristic curves. Am J Clin Pathol 2001;116(4):477–482. 11. Redman R, Zalaznick H, Mazzaferri EL, Massoll NA. The impact of assessing specimen adequacy and number of needle passes for fine-needle aspiration biopsy of thyroid nodules. Thyroid 2006;16(1): 55–60.
radiology.rsna.org n Radiology: Volume 274: Number 1—January 2015
ULTRASONOGRAPHY: Malignancy Risk Stratification in Thyroid Nodules
12. Jaume JC, Chen H. Inadequate cytology of thyroid nodules: repeat it or live with it. Indian J Surg Oncol 2011;2(2):76–77. 13. Jing X, Wey E, Michael CW. Retrospective evaluation of instituted standard adequacy criteria for on-site adequacy assessment of thyroid fine-needle aspiration. Diagn Cytopathol 2011;39(6):391–394. 14. Ravetto C, Colombo L, Dottorini ME. Usefulness of fine-needle aspiration in the diagnosis of thyroid carcinoma: a retrospective study in 37,895 patients. Cancer 2000;90(6):357–363. 15. Yoon JH, Moon HJ, Kim EK, Kwak JY. Inadequate cytology in thyroid nodules: should
Moon et al
we repeat aspiration or follow-up? Ann Surg Oncol 2011;18(5):1282–1289.
ble solid nodules of the thyroid. AJR Am J Roentgenol 2002;178(3):687–691.
16. Kwak JY, Han KH, Yoon JH, et al. Thyroid imaging reporting and data system for US features of nodules: a step in establishing better stratification of cancer risk. Radiology 2011;260(3):892–899.
19. Moon HJ, Kwak JY, Kim MJ, Son EJ, Kim EK. Can vascularity at power Doppler US help predict thyroid malignancy? Radiology 2010;255(1):260–269.
17. Kwak JY, Jung I, Baek JH, et al. Image reporting and characterization system for ultrasound features of thyroid nodules: multicentric Korean retrospective study. Korean J Radiol 2013;14(1):110–117. 18. Kim EK, Park CS, Chung WY, et al. New sonographic criteria for recommending fine-needle aspiration biopsy of nonpalpa-
Radiology: Volume 274: Number 1—January 2015 n radiology.rsna.org
20. Lee MJ, Kim EK, Kwak JY, Kim MJ. Partially cystic thyroid nodules on ultrasound: probability of malignancy and sonographic differentiation. Thyroid 2009;19(4):341–346. 21. Bellantone R, Lombardi CP, Raffaelli M, et al. Management of cystic or predominantly cystic thyroid nodules: the role of ultrasound-guided fine-needle aspiration biopsy. Thyroid 2004;14(1):43–47.
295
