Assessment of response of brain metastases to radiotherapy by PET imaging of apoptosis with 18F-ML-10
- First Online:
- Cite this article as:
- Allen, A.M., Ben-Ami, M., Reshef, A. et al. Eur J Nucl Med Mol Imaging (2012) 39: 1400. doi:10.1007/s00259-012-2150-8
- 543 Views
Early assessment of tumor response to therapy is vital for treatment optimization for the individual cancer patient. Induction of apoptosis is an early and nearly universal effect of anticancer therapies. The purpose of this study was to assess the performance of 18F-ML-10, a novel PET radiotracer for apoptosis, as a tool for the early detection of response of brain metastases to whole-brain radiation therapy (WBRT).
Materials and methods
Ten patients with brain metastases treated with WBRT at 30 Gy in ten daily fractions were enrolled in this trial. Each patient underwent two 18F-ML-10 PET scans, one prior to the radiation therapy (baseline scan), and the second after nine or ten fractions of radiotherapy (follow-up scan). MRI was performed at 6–8 weeks following completion of the radiation therapy. Early treatment-induced changes in tumor 18F-ML-10 uptake on the PET scan were measured by voxel-based analysis, and were then evaluated by correlation analysis as predictors of the extent of later changes in tumor anatomical dimensions as seen on MRI scans 6–8 weeks after completion of therapy.
In all ten patients, all brain lesions were detected by both MRI and the 18F-ML-10 PET scan. A highly significant correlation was found between early changes on the 18F-ML-10 scan and later changes in tumor anatomical dimensions (r = 0.9).
These results support the potential of 18F-ML-10 PET as a novel tool for the early detection of response of brain metastases to WBRT.
KeywordsPositron emission tomographyResponse assessmentWhole-brain radiation therapyBrain metastasesApoptosis
Brain metastasis is the most common intracranial malignancy in adults, occurring in 10 % to 30 % of adult cancer patients . Whole-brain radiotherapy (WBRT) is part of the standard management of patients with brain metastases. WBRT is highly effective in reducing neurological symptoms and preventing disease progression in the brain. However, WBRT is often associated with side effects, ranging from acute effects, such as seizure or brain edema, to diffuse delayed damage, leading, among other things, to progressive dementia . In addition, WBRT is often just the first step in the management of brain metastases, which may include more aggressive treatments such as stereotactic radiosurgery and chemotherapy in selected patients [3–6]. Currently, the standard of care for response assessment in oncology is anatomical imaging using CT or MRI. However, the time-scale limits this mode of assessment of clinical response, since the anatomical changes in the tumor observed tend to lag behind the biological effects that take place at the cellular level. As a result, the commonly accepted time-point for response assessment following WBRT is 4–8 weeks after completion of therapy [1, 7]. Since the life expectancy of patients with brain metastases is short and response is non-uniform, the early assessment of treatment efficacy in patients with brain metastases is of paramount importance. The earlier the availability of the measure of response, the greater the likelihood that the chosen follow-up treatment will be effective .
A key process in the effect of all anticancer treatments is tumor cell kill, either through proapoptotic signaling, such as those triggered by induction of DNA damage, by inhibition of antiapoptotic activity, or through stimulation of apoptotic effectors [9–13]. Apoptosis, besides other related modes of regulated cell death, such as mitotic catastrophe, plays an early and pivotal role in this process . Therefore, imaging of apoptosis may serve as a valuable early indicator of the extent of tumor response to therapy.
18F-ML-10 (2-(5-fluoro-pentyl)-2-methyl-malonic acid) is a low molecular weight PET probe, rationally designed for selective detection of apoptosis in vivo in the clinical setting. The target of the probe is a unique set of alterations that take place at the cell membrane at the point of commitment of the cell to the apoptotic death program. This selective set of alterations comprises the following elements: (1) activation of the phospholipid scrambling system that leads to exposure on the cell surface of the acidic phospholipid phosphatidylserine; (2) consequently, there is acidification of the membrane surface that leads to monoprotonation of 18F-ML-10, changing the molecule’s conformation and its interaction with the membrane interface; (3) subsequent flip-flop of the molecule through the membrane’s hydrophobic hydrocarbon core to the inner membrane leaflet, driven by scramblase activation and the irreversible depolarization of the cell membrane during the apoptotic process; and (4) binding of the molecule to cytoplasmic proteins through electrostatic and hydrophobic interactions, augmented by the irreversible loss of cellular pH control, leading to reduction in the pH of the proteins to their isoelectric points and dehydration, characteristic of the apoptotic cell. This set of alterations acts in a concerted manner at the point of commitment of the cell to the death program, and then acts as a selective transmembrane transporter of 18F-ML-10 into apoptotic cells and a driving force for its accumulation in the cytoplasm of the apoptotic cells, and not in viable or necrotic cells . In various preclinical models, both in vitro and in vivo, ranging from models of ischemic insults to anticancer therapy in cell models, selective accumulation of ML-10, which was highly correlated with histopathological assessment of apoptosis, has been observed using well-established in vitro markers such as annexin V, caspase activation and the TUNEL assay for apoptotic DNA fragmentation . In the first human trial, a phase I pharmacokinetic and safety study , 18F-ML-10 showed a favorable biodistribution profile, with rapid clearance from nontarget tissues, remarkable stability upon systemic administration (97.5 % intact compound 150 min after tracer administration), and a good safety profile, thus making 18F-ML-10 a promising radiotracer for clinical use in patients [16–18].
Considering the unmet need for tools for early assessment of response to anticancer treatments, and the promising features of 18F-ML-10, the aim of the current study was to evaluate the potential of 18F-ML-10 as a PET imaging radiotracer for early detection of response of brain metastases to WBRT.
Materials and methods
From April 2008 to December 2008, ten patients with newly diagnosed brain metastases, scheduled to undergo WBRT, were prospectively enrolled in the study. Six patients were men with a median age of 68.5 years (range 32–77 years). The histology of the primary tumor in six of the ten patients was non-small cell lung cancer. The primary tumors in the remaining patients were small cell lung cancer, uterine sarcoma, melanoma and breast cancer. The study protocol defined eligibility for the study as the presence of at least one lesion ≥1.5 cm in diameter, as assessed by gadolinium-enhanced MRI. This minimal size of the lesion was determined by the resolution constraints of PET imaging. Patients who had previously received brain irradiation, or who had concomitant uncontrolled systemic disease were excluded from the study. The study was approved by the institutional Review Board, and all patients signed informed consent.
All patients underwent CT-based simulation on a GE Light-speed CT simulator. Treatment was planned using the Eclipse treatment planning system (Varian, Palo Alto, CA) with two opposed lateral beams using customized blocking. All patients received a total radiation dose of 30 Gy in ten fractions over 2 weeks.
Preparation of 18F-ML-10 and tracer administration to the patients
The ML-10 precursor was synthesized by Albany Molecular Research (Albany, NY), and the tracer was radiolabeled at Hadassah Hospital of the Hebrew University Radiochemistry Unit (Hadassah Medical Center, Jerusalem, Israel) utilizing an IBA 18/9 cyclotron. Purification was performed using an integrated preparative HPLC system, reaching a chemical purity of >99 % by HPLC.
18F-ML-10 was administered intravenously to each patient twice followed by a PET/CT session: at baseline, before initiation of WBRT and after administration of WBRT to a cumulative radiation dose of 27–30 Gy, i.e., nine or ten fractions.
18F-ML-10 PET image acquisition
All enrolled patients were weighed for dose calculation prior to 18F-ML-10 administration, and doses did not exceed 13.5 mCi. Each PET/CT investigation session consisted of three PET/CT scans, performed at 20–36 min (eight frames, 2 min each), 80–100 min (four frames, 5 min each) and 130–150 min (two frames, 10 min each) after tracer administration. Patients were encouraged to drink and void between scanning sessions.
Imaging was performed on a GE Discovery PET/CT scanner (GE Healthcare), with a 15.7-cm axial and 70-cm transaxial field of view. Acquisition was performed in three-dimensional brain mode. In addition, a diagnostic quality CT scan of the brain was acquired at 120 mAs, serving for both attenuation and scatter correction of the PET emission data and for visualization of the brain structures.
Conventional anatomical imaging and analysis
Each patient underwent two MRI scans, the first at baseline before WBRT and then 6–8 weeks after WBRT. Each MRI scan included gadolinium-enhanced, axial T1- and T2-weighted images, each with a slice width of 3 mm in order to match the PET/CT slices. The dimensions of each brain lesion in both the baseline and follow-up MRI examinations were defined according to the region of enhancement in the gadolinium-enhanced axial T1-weighted images.
18F-ML-10 PET image analysis
The 18F-ML-10 imaging data were analyzed using a voxel-by-voxel approach. For this purpose, the baseline ML-10 PET scan was aligned with the baseline MRI scan. Then, the follow-up PET scan was coregistered to the baseline PET scan. Data registration was applied automatically using rigid body registration algorithms (ITK, version 3.20.0; Insight Toolkit by Kitware by the National Library of Medicine; www.itk.org) composed of translations and rotations. MRI/PET registration used the Mattes Mutual Information metric while PET/PET registration used the Normalized Cross-Correlation metric. Following coregistration, the image data were resampled so when the alignments of all the data were completed, all coregistered the PET and MRI images had the same voxel size and number of slices as the original baseline MRI scan with which they were coregistered. To ensure coregistration accuracy all registrations were manually inspected for correctness via the PMOD viewing tool (version 7.4.0; www.pmod.com) and its image blending functionality.
For quantification of response, the voxel-based analysis was based on a previously developed methodology [19–21] for early assessment of response of brain tumors to therapy. This methodology was developed to address the spatial heterogeneity observed in tumor responsiveness. Accordingly, a voxel-by-voxel analysis was performed, quantifying the signal for each voxel and comparing the pretreatment signal with the posttreatment signal in the respective voxel to calculate the percentage change in signal intensity in each pair of voxels before and after treatment. The mean percentage change in all voxels was then calculated. Uptake of 18F-ML-10 by each voxel was quantified according to its standardized uptake value (SUV) and normalized to the uptake in the blood pool. Posttreatment versus pretreatment plots were generated, and voxels were sorted into one of three categories: (1) voxels showing positive uptake, defined as an increase in uptake of more than 12.5 % from baseline (biologically, these values represent cells in early apoptosis, thus manifesting increased uptake of 18F-ML-10); (2) voxels showing a negative value, defined as a decrease in uptake of more than 12.5 % (biologically, such a decline may represent, among other things, vascular shutdown secondary to endothelial apoptosis and/or involution of the dead tissue with removal of the dead apoptotic cells); and (3) voxels showing no change, defined as a change in uptake that did not exceed 12.5 % (Fig. 1). The voxel signal intensity values were plotted with the pretreatment values on the abscissa and the posttreatment values on the ordinate (Fig. 1). The concept of a cut-off value, distinguishing the three zones (increased, decreased and no change) was adopted from Moffat et al. . The specific value of 12.5 % was chosen arbitrarily for presentation after confirmation that similar high correlation coefficients were also observed following evaluation of other cut-off values, e.g., 5 %, 10 % and 20 %.
For calculation of signal-to-background ratios for each lesion the mean SUV of 18F-ML-10 in the lesion was measured, and compared to the uptake in the corresponding, contralateral, healthy cerebral hemisphere (normal brain tissue).
These mathematical analyses were performed by MATLAB and Python.
Safety assessment included recording of vital signs before each administration of 18F-ML-10, at minutes 5, 10 and 30 afterwards, and at the end of the PET/CT session. A 12-lead ECG was recorded before and after each 18F-ML-10 administration. In addition, all patients were regularly followed for safety assessment for 2 months following treatment.
The PET results for each lesion were analyzed as described above, and the total percentage of voxels of the tumor manifesting an absolute change in uptake that exceeded the predetermined cut-off value of 12.5 % was calculated. Analysis of the anatomical images (CT or MRI) comprised assessment of tumor size after treatment versus tumor size before treatment, with tumor response presented as the percentage change in tumor dimensions. Tumor dimensions were determined (1) as the product of the longest perpendicular bidimensional diameters of the tumor according to the World Health Organization (WHO) method of assessment, and (2) by volumetric measurement of lesion size. The changes in tumor dimensions, as measured 6–8 weeks after completion of therapy, were then plotted against the percentage of voxels that changed in the 18F-ML-10 PET examination during treatment. The Pearson R correlation test was then applied to evaluate potential linear correlation between the percentage of tumor voxels that manifested a treatment-induced change in uptake of 18F-ML-10 as assessed by PET, and the corresponding changes in the anatomical dimensions of the tumor as assessed by MRI.
Tumor uptake of 18F-ML-10
Signal-to-background activity ratio
Nine patients completed both the pretreatment and follow-up 18F-ML-10 PET/CT sessions. The ratio between the mean uptake of 18F-ML-10 in the lesion area (defined as the ‘signal’) and 18F-ML-10 uptake in the corresponding contralateral healthy brain tissue (defined as ‘background’) was determined. Overall, high signal-to-background ratios were observed, with ratios of 4.62 ± 2.64 (mean ± SD) for the first scan performed at 20–36 min, 6.63 ± 3.81 for the second scan performed at 80–100 min, and 8.76 ± 5.59 for the last PET/CT scan performed at 130–150 min after tracer administration. Thus, the signal-to-background ratio increased over time after tracer administration, reflecting tracer accumulation at its target over time, while being cleared concomitantly from nontarget regions. Accordingly, the lowest measured signal-to-background ratio of a lesion was 1.75, observed in the first PET/CT scan, while the highest measured ratio of 20.45 was observed in the third PET/CT scan.
Imaging of treatment-induced alterations in 18F-ML-10 tumor uptake
Voxel-based analysis of treatment-induced change in 18F-ML-10 tumor uptake
The analysis included 11 tumors in nine patients who underwent two PET/CT sessions. The mean percentage of voxels that showed a change in 18F-ML-10 in the VOI was 69.9 %, ranging from 36.3 % to 100 % (95 % CI 58.5–81.3, p < 0.0001). Scatter plots of tracer uptake after treatment versus before treatment are shown in Fig. 1 (b and c, respectively). Nonresponding tumors can be clearly differentiated from responding tumors by the fraction of voxels undergoing a treatment-related change in signal intensity, either an increase or a decrease.
Tumor response as assessed by MRI, and correlation with 18F ML-10 uptake
Tumor shrinkage (anatomical parameters) and percentage of changed voxels in response to radiation treatment
Product of two diameters on MRI
Volume on MRI
Percentage of changed voxels on PET
Importantly, as shown in Fig. 3, a highly significant correlation was observed between the change in 18F-ML-10 uptake in the tumor measured early during treatment and the subsequent change in tumor mass measured at 6–8 weeks after completion of treatment, and evaluated either by the WHO method (R = 0.92) or by the volumetric method (R = 0.91).
No drug-related adverse events were observed, and no drug-related effects were identified on any of the safety parameters evaluated during the study, including ECG, laboratory tests, vital signs and physical examination.
The primary endpoint of this study was to demonstrate for the first time in the oncological setting the safety and efficacy of 18F-ML-10, a novel radiotracer used to follow response of brain metastases to radiation therapy. None of the patients enrolled in the study experienced adverse effects related to the experimental tracer. In addition, 18F-ML-10 showed highly specific affinity for the target lesions, demonstrating good signal-to-background ratios. Finally, as shown in Fig. 3, a statistically significant correlation was observed between the 18F-ML-10 signal observed after therapy as compared to the standard of care MRI. Importantly, the information obtained by 18F-ML-10 was provided as early as day 9 of radiotherapy.
A unique integration of cellular and clinical aspects qualifies 18F-ML-10 as the first small-molecule PET probe for clinical imaging of apoptosis. At the cellular level, 18F-ML-10 was rationally designed to selectively detect apoptotic cells through recognition of the complex of features of the cell at the point of commitment to the death process, as described above. The concurrence of this set of alterations is unique to apoptosis. 18F-ML-10 is a small molecule with molecular mass of only 206 kDa. In addition to its capacity to detect apoptosis at the cellular level, 18F-ML-10 is highly stable in vivo and nontoxic, and thus can be used in patients.
The utility of 18F-ML-10 as a tracer of apoptotic cell death has been established through a series of preclinical studies in various clinically relevant models , as well in several clinical trials performed in healthy volunteers  and in patients with acute ischemic stroke (manuscript in preparation). In these studies, 18F-ML-10 demonstrated high specificity in binding to the target apoptotic cells. In a kinetic microPET study in a mouse model of neurovascular cell death in experimental cerebral stroke induced by occlusion of the middle cerebral artery, 18F-ML-10 was retained in the target infarcted tissue but was cleared from nontarget tissues and healthy parts of the brain as evaluated 24 h after occlusion of the artery . In alignment with these findings in the preclinical model, in a study performed in patients with acute ischemic cerebral stroke, confirmed by clinical presentation and CT scan [16, 18], uptake of 18F-ML-10 was found in the stroke regions of each of the patients, with accumulation and retention over time, compared to the unaffected brain regions.
The clinical utility of apoptosis imaging in detection of response to therapy has been shown by SPECT imaging using Tc-annexin-V [24, 25]. While being effective in vitro, utilization of these probes for in vivo imaging was limited by suboptimal biodistribution, a low clearance rate and potential immunogenicity. Radiolabeled caspase substrates [18, 26, 27], such as the recently introduced isatin-based small nonpeptidyl caspase inhibitors, show nonspecific chemical reactivity and unfavorable biodistribution, thus limiting their advancement into clinical trials. These considerations, together with the increasing role of PET as the leading modality for molecular imaging, emphasize the need for a new clinically compatible PET probe for apoptosis.
Regulated cell death is a prominent and early effect of radiation therapy. While one may put emphasis on mitotic catastrophe in this context, Meyn et al.  and others have provided evidence that a larger percentage of primary apoptosis following radiotherapy is likely to occur than previously thought. In addition, the process of cell degradation following mitotic catastrophe largely utilizes the same molecular pathways as primary apoptosis and therefore, even if the cell death occurs as a result of reproductive cell death, the changes in the cellular membrane which allows uptake of ML-10 still prevail .
Early assessment of tumor response to therapy has always been a highly desirable goal in oncology practice. The most widely studied tracer in this context is 18F-FDG. While it has been shown to provide prognostic information on clinical outcome following therapy, it is more limited in the early stages during or following treatment, mainly due to interference by treatment- related inflammation in addition to its desired uptake by the tumor cells, thus interfering with an earlier response evaluation [28, 29]. Furthermore, the use of FDG for assessment of the response of brain metastasis to radiation is also limited due to the baseline high metabolic activity in the brain. Its recommended use, therefore, for assessing response in the clinical setting is several weeks after completion of radiation therapy.
The arena of early stages of tumor response is more complex, comprising early induction of apoptosis of the tumor cells, apoptosis of endothelial cells, subsequent vascular involution and shut-down of tumor blood vessels, removal of dead apoptotic cells, and regression of the extracellular matrix. Consequently, heterogeneity is an important factor in early tumor response. Thus various intensities of cell death signals and various changes from baseline levels can be found in various parts of the treated tumor, and in relation to the above complexity, the heterogeneity in these treatment-related processes is reflected in the uptake of 18F-ML-10. This spatial heterogeneity in response to therapy has also been observed in tumor diffusion studies (30), studies with radiolabeled annexin-V  and studies with 18F-FDG aiming to measure early effects of therapy.
In spite of the complexity of the early detection of tumor response in relation to later detection of anatomical shrinkage, this arena is of great clinical importance, since any information on tumor responsiveness provided at an early time-point may allow treatment optimization or the avoidance of treatment adverse effects. However, early imaging after treatment, with the associated signal heterogeneity within the tumor may require novel avenues for image analysis beyond the standard maximal or mean SUV. One of the hallmarks of such early changes is that they are optionally bidirectional, as shown for apparent diffusion coefficient (ADC) in diffusion-weighted MRI , FDG, annexin V, and also 18F-ML-10. We therefore chose to adopt the method of Galban et al.  for our analysis. While the main contributors to the heterogeneity of ADC values on MRI are response-related processes such as cell death, intratumoral edema or tissue reorganization, the factors for heterogeneity of the apoptotic signal of 18FML-18, which can lead to bidirectionality, may be vascular shut-down and involution of the vascular endothelium due to apoptosis of the endothelial cells, and the prothrombotic effects of apoptosis, as well as clearance of the dead cells from the tissue. Adopting the approach of a voxel-by-voxel analysis allowed the entire repertoire of these events, all related to cell death processes, to be captured and translated into numerical parameters reflecting the “response profile” of the individual patient. Importantly, this analysis showed predictive power by providing a highly significant correlation between the “response profiles” of the examined tumor and the subsequent tumor shrinkage observed on the MRI scan 2 months later.
Obviously, this small-scale study had limitations, the main one being the limited number of patients enrolled. In addition, in this study the follow-up scan was performed on day 9 or 10 of treatment. Exploration of earlier time-points may enable within-treatment modifications in the treatment protocol for better optimization of treatment in the individual patient.
In conclusion, this study showed 18F-ML-10 PET to be a novel tool with the potential to address the clinically important goal of early detection of tumor responsiveness to radiotherapy. Ongoing international multicenter studies in multiple tumor types are now being performed. It is hoped that 18F-ML-10 may serve as a new simple tracer for noninvasive imaging able to visualize cell death for the prediction of tumor responsiveness, and thereby assist in the transition to a more personalized approach in oncology.
This trial was financially supported by Aposense Ltd.