Predictive value of LN metastasis detected by 18F-FDG PET/CT in patients with papillary thyroid cancer receiving iodine-131 radiotherapy.

The aim of the present study was to predict the prognostic value of 18F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) in the metastatic lymph nodes (mLNs) of patients with papillary thyroid carcinoma (PTC) with a negative iodine-131 (131I) whole-body scan (WBS). The present retrospective study included 32 patients with PTC undergoing standard surgery and radioiodine treatment. All patients received 18F-FDG PET/CT imaging prior to and following therapy. All mLNs were divided into an effective treatment group (group A) and ineffective treatment group (group B) based on the PET Response Criteria in Solid Tumors 1.0 guidelines. All the patients were followed up for ≥9 months. A significant difference was identified in the peak standardized uptake value (SULpeak) between group B (7.85±3.20) and group A (5.36±2.19). A cut-off value of 5.85 was used to distinguish ineffective treatment of lesions from mLNs receiving radioactive ablation based on receiver operating characteristic (ROC) curve analysis with an area under the ROC curve of 0.755. Patients with a high SULpeak (P=0.003) and extrathyroidal extension (P=0.030), confirmed by pathology, more frequently exhibited a poor prognosis. In conclusion, tracer uptake of 18F-FDG for cervical metastatic nodes was revealed as a predictor for the clinical outcome of patients with PTC treated with radioiodine therapy. The present results also indicated that high SULpeak and extrathyroidal extension are poor predictors for patients with mLNs receiving 131I therapy.

Introduction

Papillary thyroid carcinoma (PTC) is the most common type of thyroid tumor and despite developments in imaging techniques, the incidence rate of PTC has increased over the last decade (1). Locoregional lymph node (LN) metastasis is typically one of the first steps in the progression of PTC to distant metastasis from the thyroid (2,3). According to a previous report from the National Comprehensive Cancer Network, the LN metastasis rate in PTC is 50–80% (4).

The most effective therapeutic strategies for well-differentiated thyroid carcinoma are thyroid surgery and iodine-131 (131I) radiotherapy (5). 131I radiotherapy is not only appropriate for primary tumors but also for LN and distant metastasis (6). Kim et al (7) reported that for lesions with high serum thyroglobulin (Tg), a negative cervical sonography and an 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) scan exhibit limited therapeutic effect with 131I therapy. Positron emission tomography and computed tomography (PET/CT) has been applied for the evaluation and monitoring of PTC and nodal metastasis, particularly for patients with elevated serum Tg levels but produced negative findings with an iodine-131 (131I) whole-body scan (WBS) (8). Previously, small LNs have been detected with PET/CT, which are easily misdiagnosed by 131I-WBS due to a low resolution and the presence of remnant thyroid tissue (9). To the best of our knowledge, the association between metastatic LNs (mLNs) and the effect of 131I therapy has not yet been reported. The aim of the present retrospective study was to evaluate the clinical efficiency of 18F-FDG PET/CT in predicting the therapeutic response of mLNs in patients with PTC following 131I therapy.

Patients and methods

Patients

The present retrospective study was approved by Ethics Committee of Xinhua Hospital Affiliated to Shanghai Jiaotong University School of Medicine and the requirement for informed consent was waived. Between January 2012 and August 2017, 106 patients at the Nuclear Medicine Department of Xinhua Hospital Affiliated with The Shanghai Jiaotong University School of Medicine (Shanghai, China) were assessed based on the following criteria: i) All patients underwent a near-total or a total thyroidectomy; ii) patients received a PET/CT examination prior to their first and following their second or later 131I radioiodine ablation treatment; iii) patients with LN metastasis were identified by fine-needle aspiration biopsy (FNAB), postoperative pathology and other clinical methods, including neck ultrasonography, CT, 131I-WBS, and serum Tg and thyroglobulin-antibody levels. Patients were excluded if they had a history or coexistence of other metastasis, including lung or bone metastasis. In addition, patients were excluded if the LNs were not confirmed by pathology, biopsy or other clinical findings. In total, 74 patients were excluded, including 16 patients with bone metastasis, 35 patients with lung metastasis and 23 patients who did not possess appropriate evidence of LN metastasis. Therefore, 18 female and 14 male patients were enrolled in the present study.

mLNs

The mLNs involved in the present study were included according to the following criteria: i) The mLN was diagnosed by postoperative pathology or FNAB; ii) the LN had a positive 131I accumulation on the 131I WBS; iii) ultrasonographic features of malignant thyroid nodules included hypoechoic lesions, a longitudinal/transverse ratio <2, a blurred or spicular margin, microlobular contour, single or multiple microcalcifications, both internal and peripheral flow, and extrathyroidal extension or an interrupted and discontinuous echogenicity of a capsule (10); iv) a mLN on CT manifestations exhibited an unenhanced center and rim enhancement, with or without fine sand calcification and mottle calcification, or a LN with a spherical shape and minimal axial diameter (MAD) >10 mm (11) and an increase in size on follow-up CT; and v) a LN of spherical shape, peak standardized uptake value (SULpeak) >2.5 with a history of stimulated Tg level >10 ng/ml during thyroid hormone withdrawal. A total of 50 nodes were identified as mLNs.

18F-FDG PET/CT

PET/CT imaging was performed using a Biograph 64 PET/CT scanner (Siemens AG, Munich, Germany). Prior to 18F-FDG PET/CT, all patients were instructed not to eat for ≥6 h to maintain their serum glucose levels at <11.1 mmol/l prior to injection. Image acquisition started ~1 h (mean, 56±2 min; range, 50–60 min) following intravenous injection of FDG (3.7 megabecquerel/kg). The lowest possible milliampere setting on the scanner was used to acquire the CT scans for attenuation correction. The whole-body CT imaging was performed from the skull to the upper part of the thigh with 120 kV, 0.5 sec rotation, 3 mm slice thickness and 0.8 mm intervals. The PET scans were obtained immediately following the whole-body CT scan, including 5–7 bed positions (1.5 min acquisition time per bed position) over the same range as the CT scan. Prior to PET/CT imaging, the serum thyrotropin (TSH) levels of patients were maintained at 30 mU/l.

131I therapy and post-therapy 131I whole-body scintigraphic imaging

Following radioiodine ablation of the residual normal thyroid tissue, 19 patients received a second dose of therapy and five also received an additional third dose of therapy. The TSH level was measured prior to 131I administration and a minimum TSH of 30 mU/l was required. 131I-NaI was orally administered, with activities ranging between 3.7 and 7.4 gigabecquerel (100–200 mCi). The 131I WBS was conducted 3–4 days following 131I administration, using a dual-head large-field γ camera with high energy collimators. An additional single-photon emission CT/CT (SPECT/CT) scan was performed with extensive cervical iodine accumulation on the WBS scan in four patients (4/32).

Imaging analysis

The acquired images were analyzed with Siemens True D (Syngo TrueD VE13A). The tracer uptake was expressed as a standardized uptake value, which was calculated according to the following formula: Measured activity concentration (Bq/ml) × body weight (g)/injected activity (Bq). To achieve a larger region of interest, the SULpeak that corrected for lean body mass was selected instead of the widely used single-pixel maximum standardized uptake value. The maximum standardized uptake has also been demonstrated to be subjected to upward bias in low-count studies compared with the SULpeak (12). All 18F-FDG PET/CT images and 131I-WBS images were visually interpreted by two experienced nuclear physicians, and a final consensus was reached for all patients.

Statistical analysis

All quantitative data are expressed as the mean ± standard deviation. A χ2 test and t-test were used to compare the SULpeak, minimum diameter, maximum diameter, LN zone, shape and density, punctate calcification, and the longitudinal and transverse ratios of the mLNs in the effective treatment group and ineffective treatment group. To assess the predictive role of SULpeak for the therapeutic effect, receiver operating characteristic (ROC) analysis was performed in both the effective treatment group and ineffective treatment group. The cut-off value was selected as 5.85, above which the best compromise between sensitivity and specificity could be achieved. Multivariate analysis of categorical variables was conducted using the survival analysis model. All statistical analysis was performed using SPSS software (version 17.0; SPSS, Inc., Chicago, IL, USA). P<0.05 was considered to indicate a statistically significant difference.

Results

Among the 32 enrolled patients, the mean ± standard deviation age was 40.6±10.6 years (range, 24–58 years). According to the 7th Edition of The TNM staging system from The American Joint Committee on Cancer (13), 18 patients had stage I disease, 6 patients had stage II disease, 3 patients had stage III disease and 5 patients had stage IV disease. During surgery, 13 patients were identified to have extrathyroidal extension and 18 patients had residual thyroid tissue on the initial 131I WBS. In the present study, clinical serum outcomes of Tg were obtained under a TSH stimulated state <1 week prior to or following the PET/CT examination. Follow-up was performed ≥9 months following examination, with a mean follow-up time of 17.87±6.80 months (range, 9–36 months).

A total of 50 lesions were diagnosed as mLNs in 32 patients, with 1–4 lesions identified per patient. A total of ten lesions were diagnosed by histopathological findings; eight were diagnosed by positive 131I accumulation on the 131I-WBS; ten were diagnosed by neck ultrasonography and 18F-FDG PET with six lesions exhibiting a blurred margin, multiple microcalcifications, internal flow and peripheral flow, and four lesions exhibiting single microcalcification and extrathyroidal extension; and twelve lesions were diagnosed by CT and 18F-FDG PET with seven lesions exhibiting a MAD >10 mm, and four lesions demonstrating an increase in size following ≥9 months. The characteristics of the mLNs are presented in Table I.

Table I.

Characteristics of metastatic lymph nodes.

Characteristics	Mean (standard deviation), range	n (50)
Minimum axial diameter, mm	7.29 (2.42), 3.00–14.00
Maximum axial diameter, mm	9.77 (2.66), 4.70–16.00
Peak standardized uptake value	6.80 (3.06), 2.60–16.60
Lymph node zone (41)
I		0
II		14
III		9
IV		9
V		2
VI		15
VII		1
Spherical shape with solid density		34
Spherical shape with soft density		16
Longitudinal and transverse ratio (<2)
Yes		38
No		12
Punctate calcification
Yes		19
No		31

The metabolic response was determined separately for changes in the SULpeak based on the PET Response Criteria in Solid Tumors 1.0 (14). A complete response (CR) was defined as a complete resolution of abnormal 18F-FDG uptake within measurable target lesions, such that the uptake was equal to or less than the mean healthy-liver activity on the follow-up examination. A partial response (PR) was defined as a decrease of >30% in any baseline PET parameter. Progressive disease (PD) was defined as an increase in any baseline PET parameter by >30% between the baseline and follow-up, or by the appearance of a new 18F-FDG-avid lesion. Stable disease (SD) was defined as non-PD, non-PR and non-CR. All mLNs were divided into the following two groups: An effective treatment group (group A) and an ineffective treatment group (group B). The lesions in group A achieved either CR or PR, while the lesions in the group B achieved either SD or PD.

The standard uptake value of lean body mass (SUL) of the 29 mLNs in group B (mean, 7.85±3.20; range, 3.00–16.60) was significantly higher compared with that of the 21 mLNs in group A (mean, 5.36±2.19; range, 2.60–12.80; P=0.027). A cut-off value of 5.85 was used to distinguish the ineffective treatment lesions from the mLNs receiving radioactive ablation based on the ROC curve analysis, with an area under the ROC curve (AUC) of 0.755. The accuracy of predicting the therapeutic effect of group A and group B was 76.19 and 72.4%, respectively (Fig. 1). Furthermore, a significant difference was identified in the shape and density between the two groups (P=0.044; Table II). However, the minimum diameter, maximum diameter, LN zone, punctate calcification, and longitudinal and transverse ratios (<2) demonstrated no significant difference between the two groups (P>0.05; Table II).

Figure 1.

ROC curve of peak standardized uptake value to predict the therapeutic effect of lymph node metastasis (n=42). ROC, receiver operating characteristic.

Table II.

Clinical characteristics of metastatic lymph nodes in group A and group B prior to therapy.

Characteristics	Group A	Group B	P-value
Number	21	29	0.422
Minimum diameter, mm			0.338
Mean (SD)	7.02 (2.16)	7.48 (2.61)
Range	4.00–12.00	3.00–14.00
Maximum diameter, mm			0.457
Mean (SD)	9.67 (2.40)	9.84 (2.86)
Range	4.70–14.00	5.00–16.00
Peak standardized uptake value			0.027
Mean (SD)	5.36 (2.19)	7.85 (3.20)
Range	2.60–12.80	3.00–16.6
Lymph node zone, n (41)			0.416
VI	5	10
Non-VI	16	19
Shape and density, n			0.044
Spherical shape with solid density	10	19
Spherical shape with soft density	8	5
Punctate calcification, n			0.242
Yes	6	13
No	15	16
Longitudinal and transverse ratios (<2), n			0.314
Yes	14	23
No	7	6

SD, standard deviation.

The effective treatment of mLNs following 131I radiotherapy was independently associated with SULpeak and extrathyroidal extension. The sex, stage, size and location of LN, Tg level, spherical shape with solid density, LN zone, punctate calcification, longitudinal and transverse ratios (<2), and residual thyriod tissue prior to therapy were not identified as risk factors for 131I therapy (Table III).

Table III.

Multivariate analysis of prognostic factors for patients with papillary thyroid cancer.

Characteristics	Odds ratio (95% confidence interval)	P-value
Standardized uptake value	1.194 (1.061–1.345)	0.003
Extrathyroidal extension	0.436 (0.206–0.923)	0.030
Age	–	0.268
Thyroglobulin	–	0.635
Sex	–	0.785
Spherical shape with solid density	–	0.874
Minimal axial diameter	–	0.300
Maximum axial diameter	–	0.510
In the central area	–	0.105
Punctate calcification	–	0.609
Longitudinal and transverse ratios (<2)	–	0.677
Residual thyroid tissue	–	0.274

In this study, seven mLNs were found positive on both 131I WBS and PET/CT image. During follow-up, it was identified that three LNs progressed into PD following radiotherapy, while the uptake of 131I WBS decreased (Fig. 3) and four nodes in the SD group demonstrated no significant changes in 18F-FDG PET/CT and a decreased uptake on the 131I-WBS (Fig. 4).

Figure 3.

A 58-year-old female patient with papillary thyroid carcinoma with a mLN. (A) LN was detected on the left side of the neck by CT with irregular shape and high density. (B) The standardized uptake values of the lesion on the PET/CT image before (upper image) and after (lower image) 131I therapy were 4.8 and 15.4, respectively. 131I WBS image revealed intermediate 131I accumulations. (C) Following 6 months, the patient received 150 mCi 131I therapy and WBS imaging demonstrated no uptake in the cervix. The lesion was defined as progressive disease. mLN, metastatic lymph node; LN, lymph node; PET, positron emission tomography; CT, computed tomography; 131I, iodine-131; WBS, whole-body scan.

Figure 4.

A 56-year-old female patient with papillary thyroid carcinoma with a mLN. CT revealed a LN with spherical shape and microcalcification in the right side of the neck. (A) The SULs of the LN (upper image) prior to and (lower image) following therapy were 47.2 and 47.6, respectively (the cross point). (B) First 131I whole-body scan image revealed high 131I accumulations in the cervix (red arrow). (C) Uptake of 131I reduced following the second therapy, however the SUL was markedly higher (the red arrow). The lesion was defined as stable disease. SUL, standardized uptake value; mLN, metastatic lymph node; LN, lymph node; CT, computed tomography; 131I, iodine-131.

Discussion

LN metastasis is common in patients with PTC (15). A previous study based on a Surveillance, Epidemiology, and End Results database with 14 years of follow-up demonstrated that an mLN can serve as a specific independent predictor for PTC, and the total survival rate of patients with mLNs was lower compared with patients without mLNs (hazard ratio, 12.597) (16). 18F-FDG PET/CT is useful for detecting the sites of metastatic disease and recurrence, which appear on radio-iodine images (17). In the present study, all 18F-FDG PET/CT acquisitions were performed under a TSH stimulated state and FDG uptake was positive in all patients with PTC with mLNs, while 131I accumulation was identified in eight lesions from 6 patients on the first WBS. In addition, 42 mLNs demonstrated a minimum diameter ≤1 cm and 15 patients with a total of 27 LNs exhibited residual thyroid tissue and exhibited star emission in the thyroid bed area on 131I WBS, which may explain why the majority of the LNs did not demonstrate positive 131I on the initial WBS. Therefore, the present study attempted to use PET/CT imaging to predict the therapeutic effect of mLNs receiving 131I radiotherapy.

At present, the most widely used imaging technologies for diagnosis of thyroid cacinoma with mLNs include ultrasonography, magnetic resonance imaging (MRI), CT, SPECT and PET/CT (18–22). Pathology is the gold standard for diagnosing LN metastasis of thyroid carcinoma (23). However, certain mLNs are located deep in the neck, which presents difficulties for diagnosis. Imaging examination serves an important role in the diagnosis and therapeutic evaluation of thyroid cancer LN metastasis. Mulla and Schulte (24) reported that ultrasonography in the diagnosis of cervical LN metastases exhibits a sensitivity of 63% and specificity of 93%. However, ultrasonography possesses limitations for the diagnosis of central compartment lymph nodal metastases, as certain LNs are located in deep regions, the trachea and surrounding structures (25). 131I SPECT/CT has been used to evaluate patients with differentiated thyroid cancer (DTC); however, residual thyroid tissue following thyroidectomy affects the uptake of 131I and imaging of mLNs (26). Compared with CT, which provides low-resolution images for soft tissue, MRI exhibits distinct advantages for the diagnosis of metastatic thyroid carcinoma, including high-resolution determination of soft tissue, no radiation exposure and a multi-sequence, multi-parameter imaging capability (27). For detection of cervical nodal metastasis in patients with DTC, MRI demonstrates a high sensitivity, accuracy, specificity and positive predictive value (19); however, MRI exhibits a low specificity and negative predictive value for the detection of lateral neck lymph nodal metastasis (28). PET/CT is useful in localizing metastasis that does not accumulate iodine in patients with elevated Tg levels (29) and has been demonstrated to have a specificity ≤90% (30). In the present study, mLNs were diagnosed by pathology, ultrasonography, CT, SPECT, PET/CT and clinical follow up. However, MRI was not analyzed in this study due to the limited number of patients examined.

The present study identified that the SULpeak of group A was significantly lower compared with that of group B prior to therapy. FDG is highly concentrated in cells with high glycolytic rates, including poorly differentiated, proliferating thyroid cancer cells (31). The combination of a low iodine concentration and a high FDG concentration has been considered to be an indicator of poorly differentiated or dedifferentiated thyroid cancer cells (32), which typically indicates a poor prognosis. Therefore, a low uptake of 18F-FDG in LNs indicates low tumor activity and a good prognosis for patients with PTC with mLNs. Sun et al (33) demonstrated that tumor/normal tissue (T/NT) of cervical LN metastases in 18F-FDG dual-head coincidence imaging is associated with the efficacy of radioiodine therapy; a high T/NT indicates a poor therapeutic effect. Based on the present results, the cut-off value of mLNs, involved in SUL with or without 131I accumulation, was >5.85 on the PET/CT image, which demonstrated improved sensitivity and specificity for predicting a poor prognosis following 131I ablation.

The present results also demonstrated that a high SULpeak of mLNs and extrathyroidal extension following thyroid surgery were significant poor prognostic factors for patients receiving therapy. A number of previous studies have demonstrated that the 131I WBS is negative for dedifferentiated PTC lesions that lack the ability to take up iodine and these dedifferentiated PTC lesions are 18F-FDG-avid (32,34,35). The present study identified that certain mLNs demonstrated positive results in both imaging examinations; however, the therapeutic effect of these mLNs has rarely been reported. In the present study, among the eight positive mLNs revealed by 131I-WBS, two maintained SD, one achieved CR and the other five progressed to PD. A possible reason for this poor prognosis may be that the LN metastases had a lack of oxygen or were poorly differentiated, which reduced the effectiveness of 131I and led to a poorer curative effect. Of the 42 negative 131I WBS, ten LNs achieved CR, 11 reached a PR, 16 remained in SD and five developed into PD. Only eight LNs possessed a minimum diameter ≥10 mm and 27 LNs displayed a star emission image in the thyroid bed area on the 131I WBS, which may have led to the majority of the LNs not being 131I-positive on the initial WBS. During subsequent therapy, 131I accumulation was observed in two mLNs, which were negative on the first 131I WBS and had SULpeaks, of 5.5 and 5.3, respectively on the PET/CT prior to therapy. Finally, both LNs achieved CR (Fig. 2).

Figure 2.

A 24-year-old male patient with papillary thyroid carcinoma with mLNs. The LNs detected by CT exhibited a spherical shape and a homogeneous density, which revealed hypermetabolic LNs in the (A) left side of the neck and in the (B) right side of the neck. The standardized uptake values of the LNs on the PET/CT images before (upper images) and after (lower images) 131I therapy were 5.7, 3.5, and 3.0, 1.1, respectively. (C) 131I whole-body scan image revealed no 131I accumulation in the two mLNs. The lesions achieved a partial response (LN in the left neck) and complete response (LN in the right neck), respectively. mLN, metastatic lymph node; LN, lymph node; PET, positron emission tomography; CT, computed tomography; 131I, iodine-131.

A number of previous studies have confirmed that macroscopic LN metastasis is a predictor that is specifically associated with PTC persistence or recurrence (36–40). Wu et al (41) investigated prognostic factors of 131I treatment for cervical LN metastases of PTC and suggested that the therapeutic effectiveness was associated with the size of the mLNs. However, the present results revealed that patients with a high SULpeak and a history of extrathyroidal extension exhibited a poorer therapeutic effect following 131I ablation. mLNs with a SULpeak >5.85 could be considered as poor prognostic factors for therapy with or without a positive 131I result. Furthermore, seven mLNs from 4 patients with PTC were both positive on 131I WBS and PET/CT, which demonstrated no remission following 131I therapy. These lesions may be associated with more aggressive characteristics and 131I therapy appears to have little to no effect on the viability of such lesions (42).

The present study had a number of limitations. Firstly, the number of subjects was small as only 50 mLNs were involved from 32 patients with PTC, which may lead to missing 131I on SPECT imaging due to a low resolution. In addition, the pathological characteristics of the primary tumor and LNs were not completely analyzed due to the limited number of samples. Furthermore, the present study was retrospective in nature and excluded certain imaging methods that could additionally be performed, including ultrasound, CT and MRI. Therefore, future studies should aim to compare the predictive value of various imaging examinations for the prognosis of thyroid carcinoma with LN metastases.

In conclusion, a high SULpeak and spherical shape with a solid density of mLNs were identified as risk predictors for the clinical outcome of patients with PTC treated with radioiodine therapy, regardless of whether 131I dose has accumulated on those nodes. When the SULpeak was >5.85, a poor therapeutic effect was revealed for mLNs. Therefore, patients with a high SULpeak of LNs and extrathyroidal extension may be considered to exhibit a poor prognosis. By introducing associated factors of mLNs on the predictive value of PET/CET, these may exhibit a useful role in the management of PTC. Further studies should focus on the detection of LNs with two positive images to identify the association between mLNs and the clinical outcome following 131I therapy.