Introduction
The lungs are the distant organs that most frequently have metastases from differentiated thyroid cancer (DTC). For these patients, 131I therapy has become the main treatment, especially in patients with 131I-avid (functioning) pulmonary metastases (PMs) [1]. The results of 131I-whole body scans (131I-WBS) and serum thyroglobulin (Tg) values are usually used to determine whether DTC patients need to receive 131I treatment. However, not all 131I-avid PMs have a good response to 131I therapy, and more than 10% of them develope into refractory iodine diseases [2]. In this way, it is more likely to cause the 131I-avid patient to receive overtreatment. Therefore, it is particularly important to screen out 131I-avid patients who are not sensitive to 131I treatment and find new indicators that predict the efficacy. Nowadays, with the popularity of [18F]fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography ([18F]FDG PET/CT), its application in DTC patients has also increased. The accumulation of FDG in malignant tumours, to a certain extent, reflects the degree of differentiation of the tissue [3, 4], so [18F]FDG PET/CT imaging can be used to predict the effect of 131I treatment of patients with 131I-avid PMs from DTC and the prognosis of them. Moreover, the status of FDG uptake in different metastatic lesions could be different even in an individual [5–7], and these changes of morphology and metabolism are closely related to the efficacy of 131I treatment and prognosis of patients. Therefore, in this study, we assessed the value of [18F]FDG PET/CT for 131I-avid PMs from DTC and observed the changes in the PMs based on per-patient and per-lesion analyses.
Material and methods
Patients
Data from 132 patients who were diagnosed with PMs and were treated with 131I between 2011 and 2018. Among them, 42 patients met the following criteria: (a) [18F]FDG PET/CT before 131I treatment for PMs; (b) PMs were positive for iodine uptake; (c) more than one course of 131I treatment after the diagnosis of PMs; (d) only measurable soft tissue components on CT, as defined by the Response Evaluation Criteria in Solid Tumours (RECIST, version 1.1) [8]. The Xin Hua Hospital Review Board approved this retrospective study.
The diagnosis of PMs was established according to one of the following criteria: (a) the lung lesion was histologically proven; and (b) 131I uptake on more than one 131I-WBS with elevated thyroid stimulating hormone (TSH) and increased Tg levels.
Therapeutic approach and follow-up schedule
All patients were instructed to follow a low-iodine diet for at least 3-4 weeks before 131I treatment. TSH levels were 85.03 ± 35.37 µIU/mL after stopping levothyroxine (L-T4) for 3–4 weeks. L-T4 therapy was administered 72 h after 131I treatment.
Adult patients with 131I-avid PMs of DTC were treated with a high-activity dose of 131I every 3–12 months. For children aged 10–18 years old, 4.625–7.4 GBq 131I was administered, and for children aged 5–10 years old, 2.775–4.44 GBq 131I was administered every 6–12 months. The cumulative activity of 131I ranged from 3.70 to 75.85 GBq. The number of 131I therapies ranged from 2–15 cycles (mean 4.5 cycles). The mean follow-up period was 44.7 ± 16.0 months.
Criteria of remission
Tumour size evaluation on anatomical imaging
The CT images of [18F]FDG PET/CT were obtained with a 3-mm slice thickness and reconstructed with a 1-mm slice thickness starting from the apex of the lungs. All CT images were obtained with the patient in the supine position. The CT images were reviewed in consensus by 2 radiologists who were blinded to the [18F]FDG PET results and clinical follow-up data.
The CT responses were assessed using Response Evaluation Criteria in Solid Tumours (RECIST, version 1.1) as follows: (i) complete response (CR), disappearance of all lesions; (ii) partial response (PR), ≥ 30% decrease in the sum of lesion diameters, taking the baseline sum of diameters as the reference; (iii) progressive disease (PD), ≥ 20% increase in the sum of lesion diameters or appearance of ≥ 1 new lesion; and (iv) stable disease (SD), neither sufficient shrinkage to qualify for PR nor sufficient increase to qualify for PD.
CR, PR, and SD were considered good responses to 131I therapy in this study.
Tg evaluation
Tg and anti-thyroglobulin antibody (TgAb) levels were obtained before 131I administration using a time-resolved immunofluorometric assay (Anytest, Symbio Lifescience Co., Ltd., Shanghai, China). After all courses of 131I therapy were administered, we compared the Tg levels of each treatment and at the last follow-up, and classified them into 3 categories [9]: i) effective: a reduction of > 25% in Tg levels; ii); stable: decreased or increased Tg by < 25% and iii) progression: Tg increased by > 25%. Effective and stable were considered good responses to 131I therapy in this study.
Images acquisition and analysis
[18F]FDG PET/CT imaging
After 3–4 weeks of thyroid hormone withdrawal (THW), patients with PMs were admitted to our department. On the 1st day after admission, [18F]FDG PET/CT scans together with other conventional assessments, including physical examination, serum TSH, serum-stimulated Tg, and serum TgAb, were performed. On average, the patients’ TSH was 85.03 ± 35.37 µIU/mL when the scans were performed. 131I treatment was performed on the 2nd day after admission. A 131I post-therapy scan was acquired 3 days after 131I oral administration.
Per-lesion imaging analysis
For each patient, a maximum of 5 lesions with [18F]FDG uptake were studied. The lesions had to be measurable on the CT scan of the [18F]FDG PET/CT. The maximum standardized uptake value (SUVmax) of each lesion (SUVmax/lesion) was measured by using a volume of interest with a standardized uptake value (SUV) expressed using the most commonly used definition of SUV (g/mL) = tissue activity (Bq/mL)/ (injected activity [Bq]/ body weight [g]). The [18F]FDG metabolic tumour volume of each lesion (MTV/lesion), representing the volume measured in the volume of interest, was determined using margin thresholds set at 40% of the maximum SUV (SUVmax). Total lesion glycolysis (TLG/lesion) represents the [18F]FDG metabolism in a given lesion and is obtained by multiplying the SUVmean by MTV.
Per-patient imaging analysis
The SUVmax/patient represents the highest SUVmax of all lesions in a given patient. The MTV of each patient (MTV/patient) represents the volume of all lesions with [18F]FDG uptake for a given patient and is calculated by adding the metabolic tumour volumes of all lesions present in that patient. The TLG of each patient (TLG/patient) represents the sum of the [18F]FDG metabolism of all lesions in a given patient.
Statistical analysis
SPSS version 22.0 was used for statistical analyses. Continuous data are expressed as the mean ± standard deviation; categorical data are presented as frequency and percentage. Continuous data were analysed using independent samples t-tests and rank tests, and categorical data were analysed using Pearson`s chi-square test. All the factors that may have affected Tg and anatomical imaging of the PMs were analysed by univariate analysis and confirmed by the chi-square test. Logistic regression was performed for multifactor analysis. Spearman correlation and Pearson correlation were used to detect the correlations between categorical variables and continuous variables. Progression-free survival (PFS), as measured by the time between the date of the diagnosis of PMs and the date of disease progression according to RECIST, version 1.1, was the primary endpoint of this study. The effect of different variables on PFS was estimated by Kaplan-Meier survival analysis. A p value of less than 0.05 was considered statistically significant.
Results
Patient characteristics
The patients’ characteristics are listed in Table 1. A total of 42 patients who had 131I-avid lung metastasis were enrolled out of 132 patients with PMs. The mean age of the subjects was 44.07 ± 16.00 years. Our retrospective study consisted of 16 men (16/42, 38%) and 26 women (26/42, 62%). The pathology was papillary TC in 39 cases and follicular TC in 3 cases. [18F]FDG uptake was found in 34 patients (34/42, 81%), while 8 patients (8/42, 19%) with 131I-avid PMs of DTC had negative [18F]FDG results after THW. The median cumulative activity of RAI was 26.83 GBq (range: 3.7–75.85 GBq). There was a good response to 131I therapy in 34 patients (34/42, 81%) and poor response in 8 (8/42, 19%) based on Tg. According to RECIST (version 1.1), 18/42 (43%) patients showed PR, while 17/42 (40%) patients had SD, 3/42 (7%) had CR, and 4/42 (10%) had PD; good responses included CR, PR, and SD.
Factors |
No. of patients |
Positive [18F]FDG |
Negative [18F]FDG |
Age |
|||
< 55y |
31(74%) |
25 (73.5%) |
6 (75%) |
≥ 55y |
11 (26%) |
9 (26.5%) |
2 (25%) |
Gender |
|||
Male |
16 (38%) |
12 (36.3%) |
4 (50%) |
Female |
26 (62%) |
22 (64.7%) |
4 (50%) |
Histological type |
|||
PTC |
39 (93%) |
33 (97.1%) |
6 (75%) |
FTC |
3 (7%) |
1 (2.9%) |
2 (25%) |
Extent of metastases |
|||
Lung only |
31 (74%) |
24 (70.6%) |
7 (87.5%) |
Lung and other organs |
11 (26%) |
10 (29.4%) |
1 (12.5%) |
Per-patient analysis
We analysed [18F]FDG uptake in patients through SUVmax, MTV, and TLG. The median SUVmax of each patient was 1.63 (range 0.43–21.39). The median MTV/patient was 1.85 cm3 (range: 0.4–16.47). The median TLG/patient was 1.74 (range: 0.2–73.34). According to the therapeutic response based on Tg, univariate analyses showed that SUVmax/patient (1.63 ± 1.06, p = 0. 061), MTV/patient (1.85 ± 0.98 cm3, p = 0.217), and TLG/patient (1.74 ± 0.67, p = 0.109) were not factors that influenced the outcome of 131I treatment, and the same results were found with CT response (p = 0.493, 0.128, 0.113). We divided the PMs into 2 subgroups according to the [18F]FDG and 131I-avid results: (1) [18F]FDG-negative and 131I-positive PMs (F-I+, n = 8); and (2) simultaneous accumulation of [18F]FDG and 131I (F+/I+, n =34); however, there was no significant difference in response to 131I therapy between the 2 groups according to both the Tg levels and RECIST (version 1.1) (p = 0.306,1.000) (Tab. 2).
Factor |
RECIST v. 1.1 |
p-value |
Tg |
p-value |
|||||
|
CR |
PR |
SD |
PD |
Effectiveness |
Stabilization |
Progression |
||
[18F]FDG |
|||||||||
Positive |
3 |
17 |
11 |
3 |
1 |
25 |
1 |
8 |
0.306 |
Negative |
0 |
1 |
6 |
1 |
8 |
0 |
0 |
||
[18F]FDG |
|||||||||
Positive |
22 |
19 |
43 |
4 |
0.044 |
60 |
3 |
25 |
0.001 |
Negative |
16 |
12 |
59 |
13 |
90 |
0 |
10 |
Per-lesion analysis
A total of 188 lesions were studied. The median SUVmax/lesion was 0.94 (range 0.43-21.39). The median MTV/lesion was 0.40 cm3 (range: 0.4–16.47). The median TLG/lesion was 0.31 (range: 0.2–73.34). [18F]FDG-positive 131I-avid PMs of DTC were significantly more common in females (p = 0.001), older patients (p = 0.038), and in patients with larger tumour sizes (p = 0.000) and higher TSH (p = 0.010). According to the Tg levels, the SUVmax/lesion (0.94 ± 0.58, p = 0. 000), MTV/lesion (0.40 ± 0.20 cm3, p = 0. 016), TLG/lesion (0.31 ± 0.10, p = 0. 000), and the size of the tumour (5.30 ± 4.00 mm, p = 0. 000) had a significant influence on 131I treatment response. However, these factors did not lead to a significant difference in response to 131I therapy according to RECIST (version 1.1) (p = 0.124, 0.256, 0.273, 0.252). Logistic regression (Enter) was performed on the above factors. The model likelihood ratio test results are shown: c2 = 26.349, p = 0.000, indicating that the model is statistically significant. The only independent factor found from the regression equation was the size of the lesion, see Table 3. In the subgroup analysis, a significant difference in response to 131I therapy was found between PMs with F-I+ and F+/I+ (p = 0.044, 0.001) according to RECIST (version 1.1) and Tg levels (Tab. 2).
Factors |
B |
S.E. |
Wald |
Sig. |
Exp (B) |
SUVmax |
0.185 |
0.184 |
1.021 |
0.312 |
1.204 |
MTV |
–0.388 |
0.419 |
0.860 |
0.354 |
0.678 |
TLG |
0.085 |
0.336 |
0.064 |
0.800 |
1.089 |
Size |
0.225 |
0.101 |
4.940 |
0.026 |
1.253 |
Constant |
–3.110 |
0.576 |
29.153 |
0.000 |
0.045 |
Because the status of different metastatic lesions, including size and [18F]FDG uptake, could be different even in an individual, we observed changes in the lesions and then analysed the relationships between these changes and efficacy. According to RECIST, we usually used the sum of the diameters of all target lesions as the basis for evaluating patients’ therapeutic efficacy, but in this study, we found that the changes in size or metabolism of individual lesions was not always consistent with therapeutic efficacy based on RECIST. For example, 131I therapy showed good responses in a given patient, but an increasing metabolism of the lesion could still be observed. Accordingly, we further divided the patients into 2 groups: (1) the changes of size and metabolism in all lesions were consistent with patients’ therapeutic efficacy (group consistency, n = 31); and (2) the changes of size and metabolism in all lesions were inconsistent with patients’ therapeutic efficacy (group inconsistency, n = 11). In the consistency vs. inconsistency comparison, we found a significant difference between the 2 groups in response to 131I therapy (p = 0.003).
Survival
The median progression-free interval (PFI) of these DTC patients with PMs was 62 months (ranging from 6 to 69 months). The Kaplan-Meier survival analysis showed that there was a significant difference in survival between the consistency and inconsistency groups (p = 0.009), but no significantly differences were seen between the [18F]FDG-positive and [18F]FDG-negative groups (p = 0.966) (Fig. 1).
Discussion
According to our study, the effective rates of 131I treatment for DTC with lung metastasis was 81% and 90%, based on Tg levels and RECIST, respectively, which were higher than the levels reported in the literature [10, 11]. This is related to the fact that we only included iodine-avid PM patients. Compared with the evaluated PD based on RECIST, slightly more patients were evaluated as having inefficiency based on Tg levels. This may show that factors regarding the function of the lesion change earlier than the morphology, and it is important to find markers predicting the status of 131I uptake by metastatic lesions, which are desirable for timely changing of the therapeutic regimen. With the rapid growth of thyroid cancer morbidity [12], [18F]FDG PET/CT scans can provide a valuable diagnostic method about functional changing [13]. [18F]FDG-avid tumours tend to be more aggressive in behaviour [14–16]. On the other hand, the reproducibility of CT measurements is known to be lower than the reproducibility of [18F]FDG calculations. Therefore, [18F]FDG PET/CT is a powerful tool for assessing DTC.
In the per-patient analysis, we evaluated [18F]FDG avidity by SUVmax, MTV, and TLG [9, 16]; however, in our study, these factors had no clear significance for predicting the 131I treatment effect of 131I-avid PMs. We also did not find a correlation between SUVmax/patient and PFS. This may be related to the limited number of patients or the shorter follow-up time. And as shown in our study, 81% (34/42) of patients showed simultaneous [18F]FDG and 131I uptake. The high proportion of [18F]FDG-positive lesions demonstrated by our study may be due to all of the lung lesions that were measurable on chest CT. [18F]FDG metabolism patterns are related to the size of the lesions [17]; the larger the lesion, the higher its [18F]FDG uptake. Some of our patients also had extrapulmonary metastases, which may have more aggressive growth than PMs alone [18]. On the other hand, the high [18F]FDG uptake maybe implicate that the clinical significance of this flip-flop phenomenon has not been fully defined [19, 20].
In the per-lesion analysis, [18F]FDG uptake was related to age, sex, diameter of lesion, and TSH level. [18F]FDG PET/CT may be more useful in older patients, females, patients with larger PMs, or patients with high TSH values. The SUVmax, MTV, TLG, and the size of the tumour were proven to be significant factors for the efficiency of 131I treatment according to Tg levels. It is considered that [18F]FDG-avid tumours tend to be less differentiated and more aggressive than those with low [18F]FDG uptake [9]. In the subgroup analysis, the treatment for lesions with simultaneous [18F]FDG and 131I uptake had poor efficacy. Several reasons may account for this phenomenon: firstly, this may imply that the lesions are partially dedifferentiated, which is prone to happen during the process of metastasis or 131I treatment [21, 22]; in addition, it may also be related to the diameter of the lesion, which is positively correlated with [18F]FDG uptake. The longer the diameter of the lesion, the more 131I treatment is required [11, 23]. The above factors mean that the absorbed dose fails to eliminate the lesion completely. Some studies have shown that [18F]FDG-avid metastases of DTC with or without 131I uptake are resistant to 131I therapy [10], which is in line with our research. Therefore, [18F]FDG-avid lesions are seldom eradicated by radioiodine therapy alone, and it should be considered for close monitoring and other options, such as surgery or external radiation. However, the only independent influencing factor was the size of the lesion. This may be related to interference from other factors.
Considering the multicentricity and polyclone of DTC, the status of different metastatic lesions could be different even in an individual [5–7]. In this study, we found that the changes in size or metabolism in some lesions were not consistent with the changes in therapeutic efficacy of some patients. Although some patients can achieve CR, PR, or SD, they still have some lesions that show a tendency to progress. We can also conclude that these patients have poor prognosis through subgroup analyses. This phenomenon shows that some lesions exhibit different degrees of differentiation in a given patient. These lesions may progress and affect the patient’s response to 131I treatment. Therefore, when we evaluate the therapeutic efficiency of patients, lesion-based analyses and quantitatively assessing the data of [18F]FDG PET/CT using SUVmax, MTV, and TLG to predict 131I-avidity for metastatic DTC would be more reliable than qualitative per-patient evaluation only. The [18F]FDG uptake PMs may be resistant to 131I treatment, and these lesions may lead to a poor prognosis for the patient. Thus, tailored treatment modalities should be chosen for the lesions that have a malignant tendency, after balancing the toxicity of systemic treatment. This approach will effectively improve the patient’s response to treatment and avoid the 131I overtreatment of patients.
The limitations of our study are that it was a retrospective study with a relatively short follow-up period (less than 10 years). The number of patients was limited in the evaluation of 131I therapeutic effects in the subgroup analysis. Moreover, the partial volume effect and respiratory motion can also significantly influence the perception of [18F]FDG uptake.
Conclusion
[18F]FDG PET/CT is a powerful tool for predicting the 131I therapeutic efficiency of patients with 131I-avid PMs of DTC. Early postoperative [18F]FDG PET/CT for them may not only reflect tumour status but also reveal prognosis information. Lesions with a larger size and higher SUVmax, MTV, and TLG may have a poor response to therapy, and the only independent factor affecting treatment response is the size of the lesion. For the patients who accepted 131I treatment, lesion-based analyses and quantitative assessment of the [18F]FDG PET/CT data would be more desirable than qualitative per-patient evaluation only. The [18F]FDG uptake PMs may lead to a poor prognosis for the patient, and early focal treatment modalities may improve their life span. A limited number of patients had 131I avidity and high [18F]FDG avidity, which may suggest refractory disease.
Compliance with ethical standards
This study was approved by the Ethics Review Board of XinHua Hospital. Because of the retrospective nature and the fact that no individually identifiable or sensitive information was involved, informed consent from the patients was waived.
Consent for publication
Not applicable.
Availability of data and materials
All data generated or analysed during this study are included in this published article and its supplementary information files.
Conflict of interest
The authors declare that they have no competing interests.
Funding
This study was supported by National Natural Science Funds of China (Grant No. 81974269).
Authors’ contributions
Z.X., S.W., H.W., and H.F. conceived and designed the experiments. Z.X., C.L., and F.F. performed experiments. Z.X. and H.F. wrote the manuscript, performed analysis, and assembled the tables and figure.
Acknowledgements
Not applicable.