Combined use of ¹⁸F-FDG and ¹⁸F-FMISO in unresectable non-small cell lung cancer patients planned for radiotherapy: a dynamic PET/CT study

Abstract

Aim of this study was to evaluate and compare, by means of dynamic and static PET/CT, the distribution patterns and pharmacokinetics of fluorine-18 fluorodeoxyglucose (¹⁸F-FDG) and of fluorine-18-fluoromisonidazole (¹⁸F-FMISO) in non-small cell lung cancer (NSCLC) patients scheduled for intensity modulated radiation therapy (IMRT). Thirteen patients suffering from inoperable stage III NSCLC underwent PET/CTs with ¹⁸F-FDG and ¹⁸F-FMISO for tumor metabolism and hypoxia assessment accordingly. Evaluation of PET/CT studies was based on visual analysis, semi-quantitative (SUV) calculations and absolute quantitative estimations, after application of a two-tissue compartment model and a non-compartmental approach. ¹⁸F-FDG PET/CT revealed all thirteen primary lung tumors as sites of increased ¹⁸F-FDG uptake. Six patients demonstrated also in total 43 ¹⁸F-FDG avid metastases; these patients were excluded from radiotherapy. ¹⁸F-MISO PET/CT demonstrated 12/13 primary lung tumors with faint tracer uptake. Only one tumor was clearly ¹⁸F-FMISO avid, (SUV_average = 3.4, SUV_max = 5.0). Mean values for ¹⁸F-FDG, as derived from dPET/CT data, were SUV_average = 8.9, SUV_max = 15.1, K₁ = 0.23, k₂ = 0.53, k₃ = 0.17, k₄ = 0.02, influx = 0.05 and fractal dimension (FD) = 1.25 for the primary tumors. The respective values for ¹⁸F-FMISO were SUV_average = 1.4, SUV_max = 2.2, K₁ = 0.26, k₂ = 0.56, k₃ = 0.06, k₄ = 0.06, influx = 0.02 and FD = 1.14. No statistically significant correlation was observed between the two tracers. ¹⁸F-FDG PET/CT changed therapy management in six patients, by excluding them from planned IMRT. ¹⁸F-FMISO PET/CT revealed absence of significant tracer uptake in the majority of the ¹⁸F-FDG avid NSCLCs. Lack of correlation between the two tracers’ kinetics indicates that they reflect different molecular mechanisms and implies the discordance between increased glycolysis and hypoxia in the malignancy.

Introduction

Lung cancer is the most common cause of cancer mortality, with non-small cell lung cancer (NSCLC) accounting for almost 80% of all cases. In 2014 the estimated new cases of lung cancer in US are 224,210, while the estimated deaths attributed to the malignancy are 159,260 [1]. The treatment of choice is surgical resection of the tumor and the draining lymph nodes. This therapeutic approach applies to patients classified in stages I-IIIA of the disease [2]. However, a radical resection is possible in only 20% of all NSCLC cases [3]. In cases of advanced, unresectable NSCLC (IIIB and IV) radiotherapy, chemotherapy or their combinations represent the standard of care. Particularly in stage IIIB NSCLC, which is the population studied in the present paper, the application of chemotherapy combined with radiation therapy results in better results than radiation therapy alone [4-6]. However, the vast majority of patients presenting with stage IIIB NSCLC still demonstrate low survival rates with an anticipated 5-year survival ranging from 3% to 7% [7]. In these terms, the identification of factors that can affect treatment response is highly important; on one side, it could lead to the selection only of those candidates who would benefit most from the applied treatment and, on the other side, it could contribute to the exclusion of those patients that are not expected to benefit from certain therapeutical approaches and, thus, avoid undergoing treatments, many of which carry some serious side-effects.

Tumor hypoxia has been recognized as an adverse prognostic factor for many cancer types with a direct negative influence on treatment success [8-11]. Hypoxia limits tumor cells’ treatment response, rendering them radioresistant, and predisposes them towards metastases [12-13]. At present the methods available for assessing oxygen concentration directly in vivo are the Eppendorf electrode (Eppendorf AG, gold standard to assess tumor hypoxia) and the Oxylite probe (Oxford optronics) [11]. However, these techniques are invasive, technically demanding and operator-dependent [11,14]. Therefore, they haven’t gained wide acceptability in clinical practice. Another well-recognised factor of crucial significance in management and prognosis of NSCLC is the extent of the disease resulting in different patient staging. The most important determinants of disease stage and subsequently prognosis are the size of the tumor, lymph node infiltration and metastatic disease [2]. In particular, distant metastatic involvement automatically upgrades patients to stage IV rendering them candidates only for palliative treatment.

PET is a non-invasive modality that, after application of appropriate radiotracers, enables the evaluation of specific molecular processes like cellular metabolism, glycolytic rate, tumor hypoxia, perfusion or receptor expression in tumors. Fluorine-18 fluoromisonidazole (¹⁸F-FMISO) is an extensively evaluated PET radiotracer in visualization of tumor hypoxia that has demonstrated a correlation between oxygen measurements and tracer uptake; ¹⁸F-FMISO PET is nowadays considered a robust method in estimating the burden of hypoxia in the pO₂ range of a few mm Hg [3,15-18]. Given the pivotal role of tumor hypoxia in cancer treatment response and prognosis, ¹⁸F-FMISO PET seems to be a very promising modality in the evaluation of oncological patients. Nevertheless, despite its wide application, ¹⁸F-FMISO remains a tracer, whose role in the diagnostic work-up of patients with NSCLC is still open. On the other hand, fluorine-18 fluorodeoxyglucose (¹⁸F-FDG), the workhorse of PET imaging, is a biomarker of intracellular glucose metabolism, reflecting the tumor ‘burden’ of the malignancy. Numerous studies have highlighted the contribution of ¹⁸F-FDG PET and PET/CT in the diagnostic approach, management and prognosis of lung cancer [19-22]. Dynamic PET/CT (dPET/CT) is a modality that allows registration of tracer kinetics over time and, after application of compartment modelling, enables the extraction of values of tracer kinetic parameters, which depict specific molecular processes [23]. Moreover, a non-compartment model can be applied leading to the estimation of fractal dimension (FD), a parameter reflecting heterogeneity [24]. This information is unattainable, when the classical whole body PET/CT protocols are performed, in which data are collected only at one time point after tracer injection.

It is known that in malignant cells glycolysis increases in hypoxia, especially in conditions of acute hypoxia. However, PET studies in several tumor types have shown that this relationship is not governed by a simple, linear pattern [25]. The purpose of this study was to evaluate and compare, by means of dynamic and static PET/CT, the distribution patterns and pharmacokinetics of ¹⁸F-FDG, a biomarker of tumor glycolytic rate and viability, and of ¹⁸F-FMISO, a biomarker of tumor hypoxia, in patients suffering from inoperable stage IIIB NSCLC. We aimed to detect potential statistically significant correlation between SUVs and kinetic parameters of the two radiopharmaceuticals that would reflect an indirect correlation between the mechanisms of glycolysis and hypoxia in NSCLC. Correlation analysis was also performed between parameters of each tracer. To our knowledge, no data have been published regarding the comparison between dynamic ¹⁸F-FDG and ¹⁸F-FMISO PET/CT studies.

Materials and methods

Patients

Thirteen patients (eleven male, two female) confirmed to suffer from inoperable stage III NSCLC were enrolled in the study. The histologic data of the patients was the following: four of them had adenocarcinoma while nine suffered from squamous cell carcinoma. Their mean age was 62.2 years (Tables 1 and 2). The included patients were not diabetic. None of them had undergone chemotherapy. Patients gave written informed consent to participate in the study and to have their medical records released. The study was approved by the Ethical Committee of the University of Heidelberg and the Federal Agency for Radiation Protection (Bundesamt für Strahlenschutz).

Table 1.

Characteristics and main results derived from PET studies with ¹⁸F-FDG of the patients investigated

Patient no.	Age (years)	Sex (M/F)	Histology	Metastases (Yes/No)	SUVaverage	SUVmax	K1	k3	influx	FD
1	71	M	Adeno Ca	No	1.5	2.4	0.13	0.05	0.01	1.06
2	65	M	Adeno Ca	Yes	8.3	12.2	0.33	0.23	0.08	1.24
3	56	M	SCC	No	4.6	7.1	0.24	0.15	0.06	1.21
4	78	M	SCC	Yes	6.9	26.1	0.08	0.14	0.02	1.28
5	75	F	SCC	No	6.9	18.7	0.14	0.12	0.03	1.22
6	73	F	Adeno Ca	No	11.1	28.1	0.22	0.12	0.04	1.37
7	44	M	SCC	No	8.6	15.8	0.20	0.14	0.05	1.39
8	56	M	SCC	Yes	9.0	23.1	0.10	0.14	0.29	1.28
9	61	M	Adeno Ca	Yes	8.6	13.0	n.a.	n.a.	n.a.	n.a.
10	67	M	SCC	Yes	10.3	13.3	0.36	0.37	0.08	1.28
11	70	M	SCC	Yes	14.9	21.4	0.22	0.30	0.07	1.34
12	74	M	SCC	No	10.7	20.9	0.32	0.14	0.06	1.29
13	70	M	SCC	No	3.2	5.1	0.24	0.06	0.02	1.15

The quantitative values refer to mean ¹⁸F-FDG values. Adeno Ca = adenocarcinoma; SCC = squamous cell carcinoma; n.a. = not acquired quantitative data (dynamic PET/CT study not performed due to patient discomfort).

Table 2.

Characteristics and main results derived from PET studies with ¹⁸F-FMISO of the patients investigated

Patient no.	Age (years)	Sex (M/F)	Histology	Metastases (Yes/No)	SUVaverage	SUVmax	K1	k3	influx	FD
1	71	M	Adeno Ca	No	1.0	1.3	0.16	0.09	0.03	1.16
2	65	M	Adeno Ca	Yes	2.1	2.7	0.42	0.02	0.01	1.18
3	56	M	SCC	No	1.0	1.3	0.21	0.14	0.02	1.09
4	78	M	SCC	Yes	1.3	2.0	0.35	0.04	0.02	1.32
5	75	F	SCC	No	1.0	1.9	0.31	0.03	0.01	1.11
6	73	F	Adeno Ca	No	1.6	2.8	0.20	0.04	0.01	1.14
7	44	M	SCC	No	0.9	1.8	0.05	0.10	0.02	1.11
8	56	M	SCC	Yes	1.4	2.2	0.31	0.04	0.02	1.16
9	61	M	Adeno Ca	Yes	3.4	5.1	n.a.	n.a.	n.a.	n.a.
10	67	M	SCC	Yes	0.8	1.1	0.45	0.03	0.02	0.99
11	70	M	SCC	Yes	1.7	2.6	0.27	0.05	0.03	1.17
12	74	M	SCC	No	0.9	1.5	0.19	0.07	0.02	1.04
13	70	M	SCC	No	1.5	2.1	0.25	0.08	0.04	1.17

The quantitative values refer to mean ¹⁸F-FMISO values. Adeno Ca = adenocarcinoma; SCC = squamous cell carcinoma; n.a. = not acquired quantitative data (dynamic PET/CT study not performed due to patient discomfort).

Data acquisition

The double-tracer study in each patient was completed in two consecutive days. The patients were intravenously administered with a maximum of 250 MBq ¹⁸F-FDG on the first day and respectively a maximum of 250 MBq ¹⁸F-FMISO on the second day. Data acquisition consisted of two parts for each tracer: the dynamic part (dPET/CT studies) and the static part. dPET/CT studies were performed over the thorax for 60 minutes using a 28-frame protocol (10 frames of 30 seconds, 5 frames of 60 seconds, 5 frames of 120 seconds and 8 frames of 300 seconds). Additional static images from the maxilla to the proximal legs were acquired in the ¹⁸F-FDG PET/CT studies and respectively from the maxilla to the pelvic entry in the ¹⁸F-FMISO PET/CT studies. The image duration was 2 minutes per bed position for the emission scans. A dedicated PET/CT system (Biograph mCT, 128 S, Siemens Co., Erlangen, Germany) with an axial field of view of 21.6 cm with TruePoint and TrueV, operated in a three-dimensional mode, was used for patient studies. A low-dose attenuation CT (120 kV, 30 mA) was used for the attenuation correction of the dynamic emission PET data and for image fusion. A second low-dose CT (120 kV, 30 mA) was performed after the end of the dynamic series in order to avoid patient movement. The last images (55-60 minutes post-injection) were used for semi-quantitative analysis. All PET images were attenuation-corrected and an image matrix of 400 × 400 pixels was used for iterative image reconstruction. Iterative images reconstruction was based on the ordered subset expectation maximization algorithm (OSEM) with 2 iterations and 21 subsets as well as time of flight (TOF). The reconstructed images were converted to SUV images based on the formula [26]: SUV = tissue concentration (Bq/g)/(injected dose (Bq)/body weight (g)). The SUV 55 to 60 minutes postinjection served for quantification of the ¹⁸F-FDG data.

Data analysis

Data analysis and evaluation consisted of visual analysis of the PET/CT scans, semi-quantitative evaluation based on SUV calculations, and quantitative analysis.

Qualitative analysis was based on visual assessment of the foci with enhanced tracer uptake on transaxial, coronal, and sagittal images for both PET/CT studies. The results of the two tracer PET/CTs were afterwards compared.

Semi-quantitative evaluation was based on volumes of interest (VOIs) and on the subsequent calculation of SUVs. VOIs were drawn with a 50% isocontour over foci of increased tracer uptake (primary tumor and potential metastatic lesions), reference tissue (normal lung parenchyma) and an arterial vessel (descending aorta). As previously mentioned, the global SUV was calculated according to the formula: SUV = tissue concentration (Bq/g)/(injected dose (Bq)/body weight (g)).

Quantitative evaluation of the dynamic PET/CT data was performed using a dedicated software [27,28]. Time activity curves (TACs) were created using VOIs drawn with a 50% isocontour, placed over foci demonstrating pathological enhanced uptake. A two-tissue compartment model with a blood component (V_B) was used for the evaluation of the rate constants (K₁, k₂, k₃ and k₄) for both radiotracers [29,30]. The two-tissue compartment model we applied is a modification of the one proposed by Sokoloff et al, which did not take into account the parameters k₄ and V_B [29]. This lack of k₄ and V_B however leads to different values of the parameters K₁ and k₃, since K₁ is dependent on V_B and k₃ on k₄. The model parameters were accepted when K₁, k₂, k₃ and k₄ were less than 1 and V_B exceeded zero. The unit for the rate constants K₁, k₂, k₃ and k₄ is 1/min, while V_B reflects the fraction of blood within the VOI. The global influx was also calculated, after applying the transport rates derived from the two-tissue compartment model analysis and using the formula: influx = (K₁ × k₃)/(k₂ + k₃). The input function was retrieved from the image data according to methods already reported in literature [31,32]. One problem in patients is the accurate measurement of input function, which theoretically requires arterial blood sampling. It has been shown however, that input function can be accurately enough retrieved from image data [31]. For the input function, the mean value of the VOI data from the descending aorta was used. A vessel VOI consisted of at least seven ROIs in sequential PET/CT images. The recovery coefficient is 0.65 for a diameter of 10 mm and for the system described above using 2 iterations and 21 subsets [33]. Partial volume correction was not performed due to the very limited partial volume effects; this limitation is attributed to the high resolution, the small pixel size, the application of VOIs and the use of a large arterial vessel.

Furthermore, a non-compartment model was used in order to calculate the fractal dimension (FD). FD is a parameter for assessment of heterogeneity of the tracer kinetics and was calculated using the time activity data in each individual voxel of a VOI. The values of the FD vary from 0 to 2 showing the deterministic or chaotic distribution of the tracer activity via time in a local volume defined by a VOI. A subdivision of 7 × 7 and a maximal SUV of 20 were applied for the calculation of FD [24].

Data were statistically evaluated using the STATA/SE 12.1 (StataCorp) software on an Intel Core (2· 3.06 GHz, 4 GB RAM) running with Mac OS X 10.8.4 (Apple Inc., Cupertino, CA, USA). The statistical evaluation was performed using descriptive statistics, box plots, Wilcoxon rank-sum test and Spearman’s rank correlation analysis. The results were considered significant for p less than 0.05 (p < 0.05).

Results

Qualitative analysis

¹⁸F-FDG PET/CT demonstrated enhanced tracer uptake in all thirteen primary lung tumors. The mean tumor diameter was 6.7 cm (range from 2.3 to 13 cm). In six patients it also revealed in total 43 metastatic lesions as sites of increased ¹⁸F-FDG accumulation, leading to disease stage upgrading and subsequent alteration in patient management. On the other hand, ¹⁸F-FMISO PET/CT depicted the 12/13 primary tumors with faint tracer uptake; only one tumor (adenocarcinoma) was ¹⁸F-FMISO avid (SUV_average = 3.4, SUV_max = 5.0). Moreover, ¹⁸F-FMISO failed to show enhanced uptake in any of the 43 metastatic lesions that ¹⁸F-FDG PET/CT revealed. Figure 1 demonstrates an example of a patient with NSCLC of the right lung. ¹⁸F-FDG PET/CT images showed enhanced ¹⁸F-FDG uptake in the anatomical area of the primary tumor with a mean SUV_average of 3.2. In the same patient ¹⁸F-FMISO PET/CT revealed only faint ¹⁸F-FMISO uptake in the tumor (SUV_average = 1.5). Figure 2 shows a patient with NSCLC of the segment 5 of the right lung. ¹⁸F-FDG PET/CT revealed the primary tumor and metastatic involvement of the supraclavicular lymph nodes, lymph nodes of the cervix, mediastinum and hilar region, as well as of the osseous structures of the thorax (thoracic spine and scapula). On the other hand, ¹⁸F-FMISO demonstrated the tumor with faint uptake and moreover it failed to depict the wide metastatic involvement. Treatment management was changed after the results of the ¹⁸F-FDG PET/CT scan. Figures 3 and 4 show a patient suffering from NSCLC of the left lung. The tumor is delineated by both ¹⁸F-FDG and ¹⁸F-FMISO, with respective mean SUVs of 8.6 and 3.4. Moreover, the patient demonstrated also metastases to contralateral mediastinal lymph nodes and the sternum, which were detected, however, only with ¹⁸F-FDG PET/CT and not ¹⁸F-FMISO PET/CT.

Figure 1.

A 70-year-old male stage IIIB lung cancer patient referred to our department for PET/CT evaluation. Transaxial PET/CT images at the level of the lungs reveal the tumor in segment 3 of the right lung with enhanced ¹⁸F-FDG uptake (upper row, arrow; SUV_average = 3.2, SUV_max = 5.0) but with faint ¹⁸F-FMISO accumulation (lower row; SUV_average = 1.5, SUV_max = 2.1).

Figure 2.

A 65-year-old patient suffering from stage IIIB NSCLC of the right lung (segment 5). Maximum intensity projection (MIP) images of ¹⁸F-FDG PET/CT (left) and ¹⁸F-FMISO PET/CT (right). ¹⁸F-FDG PET/CT demonstrates the primary tumor as well as multiple hypermetabolic ¹⁸F-FDG avid foci indicative of metastatic involvement of the supraclavicular lymph nodes, lymph nodes of the cervix, mediastinum and hilar region, as well as vertebrae of the thoracic spine and the scapula. On the other hand, ¹⁸F-FMISO demonstrates the tumor with faint uptake and fails to reveal any metastases.

Figure 3.

A 61-year-old patient with stage IIIB NSCLC of the left lung. Transaxial PET/CT images at the level of the lungs reveal the tumor in segment 6 of the left lung with enhanced tracer uptake both in ¹⁸F-FDG PET/CT (arrow, upper row; SUV_average = 8.6, SUV_max = 13.0) as well as in ¹⁸F-FMISO PET/CT (arrow, lower row; SUV_average = 3.4, SUV_max = 5.0).

Figure 4.

MIP images of ¹⁸F-FDG PET/CT (upper row) and ¹⁸F-FMISO PET/CT (lower row) of the same patient as in Figure 3. The patient demonstrates, apart from the primary tumor (red arrow), metastases to contralateral mediastinal lymph nodes and the sternum (yellow arrows), which are depicted, however, only with ¹⁸F-FDG PET/CT and not with ¹⁸F-FMISO PET/CT. The patient developed also ascites due to the malignancy.

Quantitative analysis

The mean values of the primary tumors’ kinetic parameters, derived after application of two-tissue compartment modelling on the ¹⁸F-FDG dPET/CT data, were the following: SUV_average = 8.9, SUV_max = 15.1, K₁ = 0.23 (1/min), k₂ = 0.53 (1/min), k₃ = 0.17 (1/min), k₄ = 0.02 (1/min), V_B = 0.09, influx = 0.05 (1/min). In comparison, the respective mean values of the lesions (only primary tumors) detected with ¹⁸F-FMISO dPET/CT were: SUV_average = 1.5, SUV_max = 2.2, K₁ = 0.26 (1/min), k₂ = 0.56 (1/min), k₃ = 0.06 (1/min), k₄ = 0.06 (1/min), V_B = 0.12, influx = 0.02 (1/min). Tables 1, 2 and 3 demonstrate analytically the results of the PET exams with both tracers. According to these results, SUV_average, SUV_max, k₃ and influx for ¹⁸F-FDG were significantly higher than for ¹⁸F-FMISO (Table 4). The results concerning the between tracers’ differences were considered significant for p less than 0.05 (p < 0.05).

Table 3.

Descriptive statistics of SUVs and kinetic parameters for ¹⁸F-FDG and ¹⁸F-FMISO in NSCLC patients. K₁, k₂, k₃, k₄ and influx expressed in (1/min). The values refer to primary tumors

Radiopharmaceutical	Parameters	Mean	Median	Minimum	Maximum
18F-FDG	SUVaverage	8.9	8.6	1.5	17.8
	SUVmax	15.1	14.6	2.4	28.1
	K1	0.23	0.23	0.02	0.46
	k2	0.53	0.59	0.37	0.67
	k3	0.17	0.15	0.05	0.37
	k4	0.02	0.01	0.00008	0.12
	VB	0.09	0.05	0.0004	0.65
	influx	0.05	0.06	0.01	0.11
	FD	1.25	1.24	1.06	1.38
18F-FMISO	SUVaverage	1.5	1.3	0.8	3.4
	SUVmax	2.2	2.0	1.1	5.1
	K1	0.26	0.26	0.05	0.45
	k2	0.56	0.58	0.21	0.99
	k3	0.06	0.05	0.02	0.14
	k4	0.06	0.06	0.02	0.13
	VB	0.12	0.09	0.002	0.59
	influx	0.02	0.02	0.01	0.04
	FD	1.14	1.15	0.99	1.32

Table 4.

Results of the Wilcoxon rank-sum test for the probability of differences between ¹⁸F-FDG and ¹⁸F-FMISO

		18F-FMISO parameters
		SUVaverage	SUVmax	K1	k3	influx	FD
18F-FDG parameters	SUVaverage	p < 0.0001*
	SUVmax		p < 0.0001*
	K1			p = 0.218
	k3				p = 0.001*
	influx					p = 0.027*
	FD						p = 0.007*

^*Significant probabilities (p < 0.05).

Fractal dimension (FD) was applied for further characterization of ¹⁸F-FDG and ¹⁸F-FMISO kinetics in primary tumors. The mean FD value was significantly higher for ¹⁸F-FDG (FD = 1.25) than ¹⁸F-FMISO (FD = 1.14), implying a higher degree of heterogeneity of tracer uptake.

As already mentioned, metastatic lesions were depicted in ¹⁸F-FDG PET/CT but not in ¹⁸F-FMISO PET/CT. The mean ¹⁸F-FDG values of these lesions were the following: SUV_average = 9.4, SUV_max = 14.6, K₁ = 0.25 (1/min), k₂ = 0.53 (1/min), k₃ = 0.19 (1/min), k₄ = 0.02 (1/min), V_B = 0.07, influx = 0.06 (1/min) and FD = 1.24.

We performed a Spearman’s rank correlation analysis between ¹⁸F-FDG and ¹⁸F-FMISO kinetics: no significant correlation was observed between the two tracers’ kinetic parameters (Table 5). Correlation analysis was also performed between SUVs and kinetic parameters for each tracer separately. In the case of ¹⁸F-FDG the most significant correlations (p < 0.05) were found between FD and SUV_average (r = 0.86), FD and SUV_max (r = 0.70), FD and influx (r = 0.40), influx and SUV_average (r = 0.67), influx and K₁ (r = 0.58), influx and k₃ (r = 0.69), K₁ and SUV_max (r = -0.40), as well as between k₃ and SUV_average (r = 0.57) (Table 6). In ¹⁸F-FMISO the most significant correlations were observed between FD and SUV_average (r = 0.78), FD and SUV_max (r = 0.63), influx and k₃ (r = 0.59), and between K₁ and k₃ (r = -0.76) (Table 7).

Table 5.

Results of the Spearman’s rank correlation analysis between quantitative and semi-quantitative (SUVs) parameters of ¹⁸F-FDG and ¹⁸F-FMISO. Values were considered significant for p < 0.05. No statistically significant correlation between the two tracers’ parameters is demonstrated

		18F-FMISO parameters
		SUVaverage	SUVmax	K1	k3	influx	FD
18F-FDG parameters	SUVaverage	0.201 (p = 0.555)	0.556 (p = 0.082)	-0.009 (p = 0.979)	-0.373 (p = 0.259)	-0.373 (p = 0.259)	-0.227 (p = 0.502)
	SUVmax	0.173 (p = 0.611)	0.547 (p = 0.082)	0.264 (p = 0.433)	-0.509 (p = 0.110)	-0.310 (p = 0.235)	0.036 (p = 0.916)
	K1	0.173 (p = 0.611)	0.082 (p = 0.811)	-0.055 (p = 0.873)	0.055 (p = 0.873)	-0.082 (p = 0.811)	-0.227 (p = 0.502)
	k3	0.292 (p = 0.384)	0.228 (p = 0.501)	0.364 (p = 0.272)	-0.118 (p = 0.729)	-0.127 (p = 0.709)	0.027 (p = 0.937)
	influx	0.228 (p = 0.501)	0.246 (p = 0.466)	0.109 (p = 0.750)	-0.118 (p = 0.729)	-0.346 (p = 0.298)	-0.264 (p = 0.433)
	FD	-0.009 (p = 0.979)	0.392 (p = 0.233)	-0.182 (p = 0.593)	-0.146 (p = 0.670)	-0.427 (p = 0.190)	-0.209 (p = 0.537)

Table 6.

Results of the Spearman’s rank correlation analysis between quantitative and semi-quantitative (SUVs) parameters of ¹⁸F-FDG. Values were considered significant for p < 0.05

	SUVaverage	SUVmax	K1	k3	VB	influx	FD
SUVaverage
SUVmax	0.6337*
K1	0.1484	-0.3974*
k3	0.5739*	0.3034	0.0580
VB	-0.0073	-0.2570	0.0647	0.0855
influx	0.6697*	0.0421	0.5781*	0.6905*	0.2473
FD	0.8596*	0.7021*	-0.0220	0.3785	0.0263	0.3968*

^*Significant correlation (p < 0.05).

Table 7.

Results of the Spearman’s rank correlation analysis between quantitative and semi-quantitative (SUVs) parameters of ¹⁸F-FMISO. Values were considered significant for p < 0.05

	SUVaverage	SUVmax	K1	k3	VB	influx	FD
SUVaverage
SUVmax	0.8667*
K1	0.1786	0.1576
k3	-0.2382	-0.4343	-0.7552*
VB	-0.2032	-0.1611	-0.0559	0.2378		-
influx	0.0490	-0.3082	-0.1678	0.5874*	0.1678
FD	0.7811*	0.6270*	0.2448	-0.1608	-0.1826	0.1259

^*Significant correlation (p < 0.05).

Discussion

Despite its low efficacy, radiation therapy combined with platinum-based chemotherapy is the standard treatment for inoperable NSCLC. This limited efficacy, however, can be ameliorated by intensifying radiation dose without causing significant toxicity [34]. Thus, the identification of factors that affect treatment outcome and predict therapy response is highly important. Tumor hypoxia is recognized as an independent factor of clinical outcome, which has been shown to increase resistance of malignant cells against radiation therapy, thus affecting the potency of the applied treatment [13,14,25]. The extent of the disease, in terms of lymph node infiltration and metastatic involvement, is also a factor of outmost significance in NSCLC patient management; the identification of previously unknown distant metastases automatically upgrades the patient to stage IV leading to radical changes in treatment and prognosis. It is therefore crucial to carefully evaluate and subsequently select those unresectable NSCLC patients, who are candidates for radiotherapy and are expected to benefit most from this treatment. In the present study, we examined a series of patients suffering from unresectable stage IIIB NSCLC scheduled for intensity modulated radiation therapy (IMRT) by means of dynamic and static PET/CT, in order to acquire information regarding tumor burden (by means of glycolytic rate), hypoxia and potential metastatic involvement. For this reason we applied two different PET radiotracers proven to reflect sufficiently these mechanisms; ¹⁸F-FDG, a glucose analogue, which serves as a biomarker of tumor viability and aggressiveness and ¹⁸F-FMISO, a thoroughly investigated tracer for visualization of tumor hypoxia.

Our results demonstrated the high efficacy of ¹⁸F-FDG PET/CT in detecting malignant disease and its actual extent; six subjects had in total 43 previously unknown, related to the primary tumor, metastases and were therefore excluded from planned IMRT. These results are in accordance with reported data regarding the potency of ¹⁸F-FDG PET and PET/CT in NSCLC and the pivotal role of the technique in the diagnostic work-up of this malignancy. ¹⁸F-FDG PET is a standard modality for staging NSCLC and results in a different disease stage than the one determined by standard methods in over half patients, leading to alterations in patient management [35]. Particularly in regard with M-staging, ¹⁸F-FDG PET demonstrates a sensitivity of 94% and a specificity of 97% [20].

On the other hand, ¹⁸F-FMISO, despite being the most widely used nitroimidazole imaging agent, is a tracer without a clear role yet in the diagnostic algorithm of NSCLC. ¹⁸F-FMISO PET/CT provides information on tumor hypoxia and thus, as an index of tissue oxygenation, can potentially aid in radiotherapy planning and disease prognosis, given the leading role of hypoxia in radiation resistance. FMISO enters the cells by passive diffusion, with its lipophilicity being a major determinant of the intracellular penetration. The tracer is intracellularly reduced by nitroreductase enzymes and becomes then trapped in cells with reduced oxygen partial pressure, resulting in tracer accumulation. On the other hand, in normally oxygenated cells, the parent compound is regenerated by reoxidation and no metabolite accumulation takes place [36,37]. The results of our study revealed a high ¹⁸F-FMISO retention (by visual evaluation) in only 1/13 primary tumors and in none of the 43 metastases (mean SUV_average = 1.5, mean SUV_max = 2.2). It is very likely, that this small tracer accumulation in the tumor area may be due to the scanning protocol followed (lack of late imaging). The lack of late imaging is the major limitation of the study (see below), since ¹⁸F-FMISO carries the inherent disadvantages of slow accumulation in the hypoxic lesions and limited normal tissue clearance, which limit its clinical use [38]. However, the mean and median ¹⁸F-FMISO SUV are in accordance with respective values derived from previous ¹⁸F-FMISO PET studies despite the lack of acquisition of images at 2 hours post injection (p.i.) or later. Eschmann et al. reported that in 14 NSCLC patients the mean SUV_average was 1.62 at 2 hours and 1.80 at 4 hours p.i.. The respective values for SUV_max were 2.09 and 2.40 [14]. In a prospective study by Cherk et al. 21 patients with suspected or biopsy-proven NSCLC underwent ¹⁸F-FDG and ¹⁸F-FMISO PET. Their results demonstrated that ¹⁸F-FMISO tumor uptake was significantly less the ¹⁸F-FDG uptake, while the mean ¹⁸F-FMISO SUV_max at 2 hours p.i. was 1.20. Moreover, no correlation was found between the two tracers’ SUV_max [39]. Further, our semi-quantitative results are similar with those derived from head and neck cancer patients, in which the highest so far experience has been gained, regarding hypoxia imaging. Thorwarth et al. performed dynamic ¹⁸F-FMISO PET scans in 15 patients and resulted in a median SUV_max of 2.25 at 4 hours p.i. [40]. Hicks et al reported a mean ¹⁸F-FMISO SUV_max of 2.5 ± 0.5 at 2 hours p.i. (range from 1.6 to 3.5) in 15 head and neck cancer patients. The authors applied a SUV_max of 1.6 as a threshold for characterizing a primary tumor as hypoxic [41]. Tachibana et al. studied a group of ten cancer patients (four with head and neck cancer, four with gastrointestinal cancers, one with lung cancer and one with uterine cancer) and concluded that mean tumor SUV_max was similar for 100 min and 180 minutes p.i. (2.33 ± 0.72 and 2.28 ± 0.61 respectively) [42].

A two-tissue compartment model was applied for evaluation of the kinetics of both tracers. The application of this model leads to the extraction of the kinetic parameters K₁, k₂, k₃, k₄, influx (K_i) and fractional blood volume (V_B), also called vessel density. In the case of ¹⁸F-FDG K₁ reflects the influx while k₂ reflects the efflux of the tracer, and k₃ represents its phosphorylation rate while k₄ its dephosphorylation rate. Influx (K_i) is derived from the equation = (K₁ × k₃)/(k 2 + k₃). The application of the two-tissue compartment model in ¹⁸F-FMISO leads also to the extraction of kinetic parameters, which however reflect different molecular mechanisms than ¹⁸F-FDG. In the case of ¹⁸F-FMISO, indices K₁ and k₂ describe the influx and efflux of the radiopharmaceutical into out of the cells, while k₃ and k₄ represent the trapping and re-oxygenation of ¹⁸F-FMISO [3]. Our results showed that SUV_average, SUV_max, k₃ and influx were significantly higher in ¹⁸F-FDG than in ¹⁸F-FMISO. Moreover, correlation analysis performed between the tracers’ kinetic parameters demonstrated no significant correlation. This result is of no surprise, since ¹⁸F-FMISO is a much more specific agent in comparison to the non-specific nature of ¹⁸F-FDG [18]. As already mentioned, hypoxia is a general factor affecting glucose metabolism and moreover ¹⁸F-FDG has been proposed to be a surrogate marker of tumor hypoxia [43]. However, it has been shown that tumor cell response to the phenomenon of hypoxia leads to differences between hypoxia (depicted with ¹⁸F-FMISO) and glycolytic rate (depicted with ¹⁸F-FDG), resulting in discordance in tracer uptake [18,25]. Our results are in accordance with these previous findings, since, on one side, they reflect an intense glycolytic rate in the vast majority of the studied tumors and on the other side, they demonstrate no significant correlation between the tracers’ kinetic parameters and SUVs.

Apart from applying a two-tissue compartment model for evaluation of tracers’ kinetics, we also performed a non-compartmental approach leading to the extraction of fractal dimension (FD), an index representative of tissue heterogeneity. This parameter is based on the box counting procedure of the chaos theory for the analysis of dPET data [24]. The basic concept of this approach is that an increased FD is indicative of a more chaotic tracer distribution, correlating to higher tissue heterogeneity. In the present study, the FD values in lesions showed a statistically significant difference, with a mean FD value of 1.25 for ¹⁸F-FDG and of 1.14 for ¹⁸F-FMISO. Furthermore, as already mentioned, both in ¹⁸F-FDG and ¹⁸F-FMISO, FD demonstrated a highly significant correlation with various tracer parameters: in ¹⁸F-FDG with SUV_average (r = 0.86), SUV_max (r = 0.70) and influx (r = 0.40) and in ¹⁸F-FMISO with SUV_average (r = 0.78) and SUV_max (r = 0.63). This high correlation between FD and SUVs is in agreement with previous results from our group, regarding ¹⁸F-FDG and ¹⁸F-NaF PET/CT studies in multiple myeloma; in this, also double-tracer, study FD from myelomatous lesions demonstrated a highly significant correlation with SUV_average both in ¹⁸F-FDG (r = 0.93) as well as in ¹⁸F-NaF (r = 0.97) PET/CT exams [44]. It is our belief that the application of this non-compartment derived parameter can aid in evaluation of complicated oncological differential diagnosis issues.

Limitations

As already mentioned, the major limitation of our study is the lack, due to practical reasons, of delayed imaging (e.g. 2 h and 4 hour p.i.) concerning the ¹⁸F-FMISO PET/CT imaging. Our studies involved dynamic acquisition for 60 minutes and then whole-body static imaging. This limitation may be important in ¹⁸F-FMISO, due to the slow kinetics of tracer retention resulting in high background activity over the first 60 minutes. However, due to the low FMISO uptake in the normal surrounding lung parenchyma, the uptake in the tumor was clearly visible, but low. Nevertheless, we tried to overcome this obstacle with the employment of time-activity curves (TACs) derived from VOIs placed over the lung tumors after application of two-tissue compartment modelling. With the exception of one tumor that showed intense ¹⁸F-FMISO accumulation, the TACs derived from all other primary tumors demonstrated a steady low increase in which the tracer is concentrated over time in the lesion (Figure 5). Based on the behaviour of the tracer during the first 60 minutes, we assume that this steady low increase of ¹⁸F-FMISO accumulation in the tumor area would not change significantly if late acquisition would be performed. This is in accordance to other studies mentioned before [14,39-42]. In contrary to the TACs derived from ¹⁸F-FMISO, the respective TACs from ¹⁸F-FDG showed a steady increase reflecting the increase of tracer accumulation during dynamic PET acquisition (Figure 6). Another limitation was the small number of patients included.

Figure 5.

Time activity curve of ¹⁸F-FMISO derived from a primary lung tumor, after application of two-tissue compartment modelling in the respective VOIs. The tracer’s increase over time in the lesion is low and remains so during the whole acquisition time (60 min).

Figure 6.

Time activity curve of ¹⁸F-FDG derived from the same lung tumor as in Figure 5, after application of two-tissue compartment modelling in the respective VOIs. The tracer’s increase is significantly higher than the respective rhythm of ¹⁸F-FMISO.

Conclusion

We performed a double-tracer PET/CT study with the agents ¹⁸F-FDG and ¹⁸F-FMISO in patients with advanced NSCLC scheduled for IMRT. In terms of clinical practice, our results confirmed the utility of ¹⁸F-FDG PET/CT in the diagnostic approach of the malignancy by excluding 6/13 (46%) patients with distant metastases from radical radiation treatment. On the other hand, imaging with ¹⁸F-FMISO PET/CT didn’t significantly aid in the management of our patient group. Only one subject demonstrated increased ¹⁸F-FMISO tumor uptake, reflecting its potential sensitivity to radiotherapy. Based on the pharmacokinetic results, the lack of correlation between ¹⁸F-FDG and ¹⁸F-FMISO implies the discordance between increased glycolysis and hypoxia in the malignancy. To our knowledge this is the first study evaluating the distribution patterns and pharmacokinetics of these two tracers in NSCLC by means of dynamic PET/CT scanning. It is an undisputable fact that further studies are warranted in order to clarify the potential role of ¹⁸F-FMISO PET/CT in the diagnostic workup and management of the malignancy as already proposed [14,45,46].

Disclosure of conflict of interest

The authors declare that they have no conflict of interest.