Standardized uptake value differences between primary and metastatic lesions in 18 F-FDG PET/CT of patients with lung cancer

Abstract

Background

FDG-PET/CT is a robust tool for staging of lung cancer, but the differences in FDG uptake between primary and metastatic lesions have not yet been well described.

Purpose

To define the potential range of standardized uptake value (SUV) differences between primary and metastatic lesions in lung cancer patients and to identify the factors responsible for these differences.

Material and Methods

FDG-PET/CT images of 75 lung cancers with 296 metastases were analyzed retrospectively. Histological types, primary locations, and metastatic sites were recorded. The average and maximum SUV (SUVavg, SUVmax) of each primary tumor and metastasis were measured, and the ratio of metastatic SUVs to primary SUVs (M/Pavg, M/Pmax), its difference from 100% (diff-M/Pavg, diff-M/Pmax), the ratio of ROI area of metastatic to primary lesions (ROI-M/P), and its difference from 100% (diff-ROI-M/P) were calculated.

Results

M/Pavg was in the range of 35.9–224.6% (mean ± SD: 97.9% ± 35.9%), while M/Pmax was in the range of 24.8–286.7% (98.1% ± 45.3%). Furthermore, values were in the range of 50–200% for M/Pavg in 280/296 lesions (94.6%) and for M/Pmax in 255/296 lesions (86.1%). M/Pavg and M/Pmax showed significant linear correlations with ROI-M/P (r = 0.62, 0.64, respectively). Multivariate analysis showed that diff-ROI-M/P had the greatest effect on diff-M/Pavg and diff-M/Pmax.

Conclusion

The SUVs of most metastatic lesions ranged from half to double those of primaries in lung cancer patients. When the SUV of a suspected metastasis is beyond the range of half to double that of the primary lung cancer, other non-metastatic lesions should be considered, while taking ROI size into account.

Keywords

Fluorodeoxyglucose (FDG)lung cancer metastasis positron emission tomography (PET)standardized uptake value (SUV)

Introduction

¹⁸F-fluorodeoxyglucose positron emission tomography and computed tomography (FDG-PET/CT) has come to be used routinely for lung cancer in clinics for decades, and it is now used for several purposes, including diagnosis, staging, therapy response evaluation, and checking for recurrence (1,2). This robust tool provides additional information over other conventional imaging techniques, since its findings depend on the glucose metabolic state of the tumor rather than morphological changes. Therefore, FDG-PET/CT is useful for accurate diagnosis of suspicious lesions that are equivocal on conventional imaging. However, in daily practice, it can also be seen that a suspicious metastatic lesion shows quite higher or lower FDG uptake compared to that of the primary lesion. In such cases, it is difficult to determine whether this uptake is acceptable for a metastatic lesion or whether it should be treated as a second pathogenic lesion, such as a concurrent other primary neoplasm or non-neoplastic lesion (3). To the best of our knowledge, the FDG uptake difference between primary and metastatic lesions on FDG-PET/CT of patients with lung cancer has not yet been well evaluated. In the present study, therefore, FDG-PET/CT findings of patients with lung cancer having clinically definitive metastatic lesions were retrospectively reviewed, the potential range of standardized uptake value (SUV) differences between primary and metastatic lesions in lung cancer patients was defined, and the factors responsible for these differences were identified.

Material and Methods

Patients

This study was conducted according to the ethical standards of the Declaration of Helsinki, and The Local Ethics Committee of the University of Fukui approved a waiver of the need for individual informed consents. From the department database, 204 consecutive FDG-PET/CT examination of patients with lung cancer (January 2009 to December 2011) were selected. Reviewing both the clinical records and the follow-up imaging examinations, 70 patients with no metastases were excluded. Then, the following inclusion criteria were used to define a definitive metastatic lesion: (i) pathologically confirmed metastatic lesion; (ii) hilar and mediastinal lymph nodes >1 cm in the short axis on CT images, and increasing on follow-up or decreasing with chemo/radiotherapy; and (iii) apparent lesions (except lymph nodes) on CT/MR images, and increasing on follow-up or decreasing with chemo-/radiotherapy. Patients with concurrent cancer or a history of other cancers were also excluded. Fifty-nine patients did not meet the study criteria and were excluded from the study. Finally, 296 metastatic lesions in 75 patients (16 women, 59 men; mean age, 68.8 years; age range, 48–84 years) were included. All patients did not receive any chemo/radiotherapy prior to the FDG-PET/CT scan. The follow-up interval fulfilling the above inclusion criteria for a definitive metastatic lesion was an average of 49.9 days (median, 44 days), with a range of 17–193 days. Histological types (adenocarcinoma, 109 metastatic lesions in 21 patients; squamous cell carcinoma, 95 in 33 patients; small cell carcinoma, 65 in 14 patients; others, 27 in 7 patients), primary-tumor locations (right upper lobe, 27 patients; right middle lobe, 0 patients; right lower lobe, 17 patients; left upper lobe, 18 patients; left lower lobe, 13 patients), and metastatic sites (lymph node, 197 metastatic lesions; bone, 62 metastatic lesions; liver, 14 metastatic lesions; pleura/lung, 15 metastatic lesions; others [adrenal, kidney, and small intestine], 8 metastatic lesions) were recorded.

PET/CT examination

FDG-PET/CT examinations were performed using a PET/CT camera (Discovery LS, GE Healthcare, Milwaukee, WI, USA). After the patients had fasted for at least 6 h, FDG-PET/CT images were obtained 50 min after the intravenous injection of 185 MBq of FDG, and CT-based attenuation corrections were performed. Acquisition and reconstruction parameters were 2-min emission per bed position, seven bed positions, two-dimensional (2D) acquisition, 50-cm axial field of view, and an ordered-subsequent expectation maximization iterative reconstruction (subsets, 14; number of iterations, 2) with 7-mm slice thickness. Finally, reconstruction images were converted to standardized uptake value (SUV) images, using the following equation:

SUV = activity at a pixel (Bq / mL) / (injection dose (Bq) / weight (g))

Image analysis

For FDG-PET/CT image analysis, single 2D regions of interest (ROIs) that were as large as possible were placed over the primary and metastatic tumors using information obtained from fused PET/CT images, and ROI area and average and maximum SUVs (SUVavg and SUVmax, respectively) were measured. Slices displaying maximum tumor size on the axial images were selected. When FDG accumulations were apparently low at the center of the tumors (primary lesion, n = 13; metastatic lesion, n = 6), representing the necrotic portions, these portions were excluded from the ROIs.

The ratio of metastatic SUVs to primary SUVs (M/Pavg [%] = SUVavg of metastasis/primary × 100, M/Pmax [%] = SUVmax of metastasis/primary × 100) was calculated, along with its difference from 100% (diff-M/Pavg = |100 – M/Pavg|, diff-M/Pmax = |100 – M/Pmax|). The ratio of ROI area of metastatic to primary lesions (ROI-M/P [%] = ROI area of metastasis/primary× 100) and its difference from 100% (diff-ROI-M/P = |100 – ROI-M/P|) were also calculated.

Statistical analysis for M/Pavg and M/Pmax against ROI-M/P was performed using Spearman’s rank correlation test. Furthermore, one-way analysis of variance with post hoc testing was performed to compare means of diff-M/Pavg and diff-M/Pmax among subgroups of histological type, primary location, and metastatic site. Finally, to identify the effect of each factor on the SUV differences between primary and metastatic lesions, multivariate analysis was also performed with multiple regression analysis using dummy variables. Values of P < 0.05 were considered significant.

Results

M/Pavg was in the range of 35.9–224.6% (mean ± SD, median: 97.9% ± 35.9%, 89.3%), while M/Pmax was in the range of 24.8–286.7% (98.1% ± 45.3%, 89.4%). Furthermore, values were in the range of 50–200% for M/Pavg in 280/296 lesions (94.6%) and for M/Pmax in 255/296 lesions (86.1%). M/Pavg and M/Pmax showed significant linear correlations with ROI-M/P (P < 0.0001; r = 0.62, 0.64, respectively) (Fig. 1). Univariate analysis identified significant differences (P < 0.05) in diff-M/Pmax between squamous versus small cell carcinoma (mean ± SD, median: 39.8 ± 26.6, 35.0 vs. 32.9 ± 32.2, 19.5, respectively) and lymph node versus liver (38.2 ± 31.1, 31.9 vs. 17.9 ± 13.6, 17.1, respectively) (Fig. 2). Other factors examined in this study did not show significant differences. For a further evaluation of these significant differences in diff-M/Pmax, the range of SUVmax at both primary and metastatic lesions is summarized in Table 1. The coefficient of variation (CV) of SUVmax was higher in metastatic lesions of squamous cell carcinoma than of small cell carcinoma (0.66 vs. 0.43, respectively), while the CV of SUVmax in the primary lesions did not show much difference (0.38 vs. 0.37, respectively). Multivariate analysis showed that diff-ROI-M/P had the greatest effect on diff-M/Pavg and diff-M/Pmax (P < 0.001, standardized partial regression coefficient = 0.48 and 0.54, corrected R²= 0.26 and 0.30, respectively), followed by the left lower lobe (standardized partial regression coefficient = 0.19) for diff-M/Pavg, and liver (standardized partial regression coefficient = –0.24) for diff-M/Pmax. A representative case showing considerably different FDG uptakes between primary and metastatic lesions is shown in Fig. 3.

Table 1.

Summary of maximum standardized uptake values (SUVmax) of primary and metastatic lesions in squamous cell carcinoma (SqCC), small cell carcinoma (Small), lymph node (LN) metastasis, and liver metastasis.

SUVmax	SqCC	Small	LN	Liver
Primary
Subjects (n)	33	14	64	5
Mean	13.80	8.03	11.98	10.41
Median	13.23	7.41	11.77	6.81
SD	5.20	2.99	4.85	5.59
CV	0.38	0.37	0.41	0.54
Min	3.72	4.54	3.72	6.06
Max	23.99	16.18	23.99	18.79
Metastatic
Lesions (n)	95	65	197	14
Mean	12.66	9.21	11.17	8.47
Median	9.51	8.04	9.37	7.68
SD	8.33	4.00	6.69	2.98
CV	0.66	0.43	0.60	0.35
Min	3.77	4.19	2.90	4.46
Max	38.04	22.88	36.82	16.58

CV, coefficient of variation; max, maximum; min, minimum; SD, standard deviation.

Discussion

FDG-PET/CT has become an essential imaging modality in clinics for the management of patients with lung cancer (1,2). For staging, metastatic lesions generally show similar FDG uptake compared to primary lesions. However, some atypical cases that show considerably higher or lower FDG uptakes in the suspected metastatic lesions are also encountered in daily practice. In the present study, FDG-PET/CT scans of lung cancer patients with definitive metastatic lesions were retrospectively reviewed, and the possible differences in the SUVs between primary and metastatic lesions were identified. The majority of SUVs of metastatic lesions (>90% for SUVavg, >85% for SUVmax) ranged from half to double those of primary SUVs. Regarding the FDG uptake difference between primary and metastatic lesions in lung cancer, Uesaka et al. previously reported the retention index (RI) of primary and metastatic lesions in lung cancer using dual time-point FDG-PET (4). The RI was calculated by the following equation in their study: RI (%) = (SUV at the delayed scan - SUV at the early scan) × 100 / SUV at the early scan. They reported that the RI of metastatic lesions was approximately 0.5–2 times the RI of the primary lesions, which was similar to the present result; the SUV difference between the primary and metastatic lesions fell in this range even in single time-point FDG-PET. This may be a simpler assessment of the FDG uptake difference between primary and metastatic lesions. When the SUV of a suspicious metastatic lesion shows higher or lower FDG uptake beyond this range, other non-metastatic lesions should be considered in daily practice.

Other significant differences were identified. Squamous cell carcinoma and lymph node metastases showed higher diff-M/Pmax than small cell carcinoma and liver metastases, respectively. The further evaluation of these differences in SUVmax showed a higher CV of SUVmax in metastatic lesions from squamous cell carcinoma than from small cell carcinoma, while a comparable CV was identified in the primary lesions. In other words, the degree of SUVmax variation does not differ in primary lesions between squamous and small cell carcinomas, but metastatic lesions from squamous cell carcinoma shows higher degrees of SUVmax variation than from small cell carcinoma. The reason for this difference is unclear, but it has been reported that squamous cell carcinoma shows a relatively higher and wider range of SUVs compared to other types of non-small cell lung cancers (5,6), and also compared to small cell carcinoma, as observed in the present study. This might contribute to a wider variety of metabolic states of metastases from squamous cell carcinoma, although the CVs of SUVmax did not differ at the primary site. Regarding the significant diff-M/Pmax difference between lymph node and liver metastases, a lower CV was identified in liver than in lymph node metastases. Although it might be possible that the metastatic location affected the variation in SUVmax, it is more likely that the small number of subjects in the liver metastasis group (5 patients) compared to the lymph node metastasis group (64 patients) might have resulted in the small variation of SUVmax in liver metastases.

The second purpose of this study was to identify the factors affecting these SUV differences between the primary and metastatic lesions. The present results clearly indicated that ROI size difference had the strongest effect on the SUV difference between primary and metastatic lesions. A relationship between tumor size and FDG uptake in patients with lung cancer has been recently reported by Stiles et al. (7). They retrospectively reviewed the FDG-PET images and absolute tumor size of 530 surgically resected cases and concluded that increasing tumor size was an independent predictor of higher SUVmax. This may probably be explained by the fact that biologically aggressive tumors tend to grow rapidly and consume more glucose. Thus, it may be reasonable that ROI size (i.e. tumor size) differences affected the SUV difference in the present study. Moreover, partial volume effects of PET imaging may be another, probably more possible, explanation for the strong effect of ROI size on the difference. The partial volume effect emerges from both limited spatial resolution and image sampling and causes underestimation of SUVs, especially in small tumors (8). Therefore, this may also explain the strong effect of ROI size observed in this study. Furthermore, since the ROIs used in this study were 2D, which could not cover all tumor slices, the measured SUVavg and SUVmax values may not correctly represent the FDG-uptakes of entire tumors. Recently, peak SUV, which is deﬁned as the average SUV within a small, ﬁxed-size, three-dimensional region of interest (ROI peak) centered on a high-uptake part of the tumor, has been suggested for quantification of FDG uptake, since this can minimize the partial volume effect (9). Using such quantitative parameters, the ROI size effect observed in the present study might decrease, eliminating the partial volume effect.

Several limitations were identified in this study. First, since respiratory-gating was not used for PET image acquisition, respiratory movement may have caused underestimation of SUVs in the primary tumors, especially those located in the lower lobes (10). However, univariate analysis of both SUVavg and SUVmax did not show any significant differences among the primary locations; thus, it is unlikely that this had a critical effect in this study. Second, the FDG-PET/CT imaging protocol used in the present study (e.g. 2D acquisition, slice thickness, relatively low dose of FDG) may have affected the quantification of FDG uptake. For instance, large pixels (thick slices) have greater partial volume effects than small pixels (8), and a new acquisition technique, time-of-flight acquisition, has been reported to overestimate SUV in oncologic patients (11). However, the parameter used in the present study was the ratio between primary and metastatic lesions, which may be less sensitive to such factors than absolute values. Third, the follow-up intervals fulfilling the inclusion criteria for a definitive metastatic lesion were relatively short in some cases (shortest 17 days), thus non-metastatic lesions showing early size change (e.g. inflammation) might be included in this study. However, since the overall follow-up periods were much longer in the most cases, such possibility is presumed to be limited in this study. Lastly, FDG-avid non-metastatic lesions, such as concurrent other neoplasms, inflammation, or sarcoidosis, were not included in this study. Thus, even if the suspicious metastatic lesion showed an uptake within the range of 0.5–2 times compared to the primary lesion, the possibility of non-metastatic lesions cannot be ruled out. From our clinical experience, similar FDG uptake to primary tumor can also be observed in non-metastatic lesions. Thus, it may be unrealistic to rule out the possibility of such lesions by the degree of FDG-uptake alone. Furthermore, previous literature also suggests that positive findings of FDG-PET alone are non-specific findings in sarcoidosis or inflammatory disease, although the clinical utility of FDG-PET for whole body screening or therapeutic effect monitoring has been reported (12,13). In such cases, other imaging findings of high-resolution CT of the lungs or contrast-enhanced CT/MRI, as well as laboratory examinations, will provide helpful information to rule out the possibility of such non-metastatic lesions.

In conclusion, the SUVs of most metastatic lesions ranged from half to double those of primaries in lung cancer patients, and the ROI size difference between two lesions was found to be the factor with the greatest effect on these differences. From the clinical perspective, when the SUV of a suspected metastasis is beyond 0.5–2 times that of the primary, other non-metastatic lesions, such as a concurrent neoplasm or non-neoplastic lesions, should be considered, while taking ROI size differences into account.

Footnotes

Conflict of interest

None declared.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

References

Cuaron

Dunphy

Rimner

. Role of FDG-PET scans in staging, response assessment, and follow-up care for non-small cell lung cancer. Front Oncol 2013; 2: 208–208.

van Tinteren

Hoekstra

Smit

. Effectiveness of positron emission tomography in the preoperative assessment of patients with suspected non-small-cell lung cancer: the PLUS multicentre randomised trial. Lancet 2002; 359: 1388–1393.

Chopra

Ford

De Noronha

. Incidental findings on positron emission tomography/CT scans performed in the investigation of lung cancer. Br J Radiol 2012; 85: e229–e237.

Uesaka

Demura

Ishizaki

. Evaluation of dual-time-point 18F-FDG PET for staging in patients with lung cancer. J Nucl Med 2008; 49: 1606–1612.

Aquino

Halpern

Kuester

. FDG-PET and CT features of non-small cell lung cancer based on tumor type. Int J Mol Med 2007; 19: 495–499.

Suzawa

Ito

Qiao

. Assessment of factors influencing FDG uptake in non-small cell lung cancer on PET/CT by investigating histological differences in expression of glucose transporters 1 and 3 and tumour size. Lung Cancer 2011; 72: 191–198.

Stiles

Nasar

Mirza

. Ratio of positron emission tomography uptake to tumor size in surgically resected non-small cell lung cancer. Ann Thorac Surg 2013; 95: 397–404.

Soret

Bacharach

Buvat

. Partial-volume effect in PET tumor imaging. J Nucl Med 2007; 48: 932–945.

Wahl

Jacene

Kasamon

. From RECIST to PERCIST: Evolving Considerations for PET response criteria in solid tumors. J Nucl Med 2009; 50(Suppl. 1): 122S–150S.

10.

Kawano

Ohtake

Inoue

. Deep-inspiration breath-hold PET/CT versus free breathing PET/CT and respiratory gating PET for reference: evaluation in 95 patients with lung cancer. Ann Nucl Med 2011; 25: 109–116.

11.

Prieto

Dominguez-Prado

Garcia-Velloso

. Impact of time-of-flight and point-spread-function in SUV quantification for oncological PET. Clin Nucl Med 2013; 38: 103–109.

12.

Glaudemans

de Vries

Galli

. The use of (18)F-FDG-PET/CT for diagnosis and treatment monitoring of inflammatory and infectious diseases. Clin Dev Immunol 2013; 2013: 623036–623036.

13.

Sobic-Saranovic

Artiko

Obradovic

. FDG PET imaging in sarcoidosis. Semin Nucl Med 2013; 43: 404–411.