Alexandr Lougovski

Asphericity of pretherapeutic tumour FDG uptake provides independent prognostic value in head-and-neck cancer

AbstractObjectiveTo propose a novel measure, namely the ‘asphericity’ (ASP), of spatial irregularity of FDG uptake in the primary tumour as a prognostic marker in head-and-neck cancer.MethodsPET/CT was performed in 52 patients (first presentation, n = 36; recurrence, n = 16). The primary tumour was segmented based on thresholding at the volume-reproducible intensity threshold after subtraction of the local background. ASP was used to characterise the deviation of the tumour’s shape from sphere symmetry. Tumour stage, tumour localisation, lymph node metastases, distant metastases, SUVmax, SUVmean, metabolic tumour volume (MTV) and total lesion glycolysis (TLG) were also considered. The association of overall (OAS) and progression-free survival (PFS) with these parameters was analysed.ResultsCox regression revealed high SUVmax [hazard ratio (HR) = 4.4/7.4], MTV (HR = 4.6/5.7), TLG (HR = 4.8/8.9) and ASP (HR = 7.8/7.4) as significant predictors with respect to PFS/OAS in case of first tumour manifestation. The combination of high MTV and ASP showed very high HRs of 22.7 for PFS and 13.2 for OAS. In case of recurrence, MTV (HR = 3.7) and the combination of MTV/ASP (HR = 4.2) were significant predictors of PFS.ConclusionsASP of pretherapeutic FDG uptake in the primary tumour improves the prediction of tumour progression in head-and-neck cancer at first tumour presentation.Key Points•Asphericity (ASP) characterises the spatial heterogeneity of FDG uptake in tumours • ASP is a promising prognostic parameter in head-and-neck cancer • ASP is useful for identification of high-risk patients with head-and-neck cancer

The PET-derived tumor-to-blood standard uptake ratio (SUR) is superior to tumor SUV as a surrogate parameter of the metabolic rate of FDG

BackgroundThe standard uptake value (SUV) approach in oncological positron emission tomography has known shortcomings, all of which affect the reliability of the SUV as a surrogate of the targeted quantity, the metabolic rate of [18F]fluorodeoxyglucose (FDG), Km. Among the shortcomings are time dependence, susceptibility to errors in scanner and dose calibration, insufficient correlation between systemic distribution volume and body weight, and, consequentially, residual inter-study variability of the arterial input function (AIF) despite SUV normalization. Especially the latter turns out to be a crucial factor adversely affecting the correlation between SUV and Km and causing inter-study variations of tumor SUVs that do not reflect actual changes of the metabolic uptake rate. In this work, we propose to replace tumor SUV by the tumor-to-blood standard uptake ratio (SUR) in order to distinctly improve the linear correlation with Km.MethodsAssuming irreversible FDG kinetics, SUR can be expected to exhibit a much better linear correlation to Km than SUV. The theoretical derivation for this prediction is given and evaluated in a group of nine patients with liver metastases of colorectal cancer for which 15 fully dynamic investigations were available and Km could thus be derived from conventional Patlak analysis.ResultsFor any fixed time point T at sufficiently late times post injection, the Patlak equation predicts a linear correlation between SUR and Km under the following assumptions: (1) approximate shape invariance (but arbitrary scale) of the AIF across scans/patients and (2) low variability of the apparent distribution volume Vr (the intercept of the Patlak Plot). This prediction - and validity of the underlying assumptions - has been verified in the investigated patient group. Replacing tumor SUVs by SURs does improve the linear correlation of the respective parameter with Km from r = 0.61 to r = 0.98.ConclusionsSUR is an easily measurable parameter that is highly correlated to Km. In this respect, it is clearly superior to SUV. Therefore, SUR should be seriously considered as a drop-in replacement for SUV-based approaches.

Increased evidence for the prognostic value of primary tumor asphericity in pretherapeutic FDG PET for risk stratification in patients with head and neck cancer

PurposeIn a previous study, we demonstrated the first evidence that the asphericity (ASP) of pretherapeutic FDG uptake in the primary tumor provides independent prognostic information in patients with head and neck cancer. The aim of this work was to confirm these results in an independent patient group examined at a different site.MethodsFDG-PET/CT was performed in 37 patients. The primary tumor was delineated by an automatic algorithm based on adaptive thresholding. For the resulting ROIs, the metabolically active part of the tumor (MTV), SUVmax, SUVmean, total lesion glycolysis (TLG) and ASP were computed. Univariate Cox regression with respect to progression free survival (PFS) and overall survival (OS) was performed. For survival analysis, patients were divided in groups of high and low risk according to the parameter cut-offs defined in our previous work. In a second step, the cut-offs were adjusted to the present data. Univariate and multivariate Cox regression was performed for the pooled data consisting of the current and the previously described patient group (N = 68). In multivariate Cox regression, clinically relevant parameters were included.ResultsUnivariate Cox regression using the previously published cut-off values revealed TLG (hazard ratio (HR) = 3) and ASP (HR = 3) as significant predictors for PFS. For OS MTV (HR = 2.7) and ASP (HR = 5.9) were significant predictors. Using the adjusted cutoffs MTV (HR = 2.9/3.3), TLG (HR = 3.1/3.3) and ASP (HR = 3.1/5.9) were prognostic for PFS/OS. In the pooled data, multivariate Cox regression revealed a significant prognostic value with respect to PFS/OS for MTV (HR = 2.3/2.1), SUVmax (HR = 2.1/2.5), TLG (HR = 3.5/3.6), and ASP (HR = 3.4/4.4).ConclusionsOur results confirm the independent prognostic value of ASP of the pretherapeutic FDG uptake in the primary tumor in patients with head and neck cancer. Moreover, these results demonstrate that ASP can be determined unambiguously across different sites.

Correction of scan time dependence of standard uptake values in oncological PET

BackgroundStandard uptake values (SUV) as well as tumor-to-blood standard uptake ratios (SUR) measured with [ 18F-]fluorodeoxyglucose (FDG) PET are time dependent. This poses a serious problem for reliable quantification since variability of scan start time relative to the time of injection is a persistent issue in clinical oncological Positron emission tomography (PET). In this work, we present a method for scan time correction of, both, SUR and SUV.MethodsAssuming irreversible FDG kinetics, SUR is linearly correlated to Km (the metabolic rate of FDG), where the slope only depends on the shape of the arterial input function (AIF) and on scan time. Considering the approximately invariant shape of the AIF, this slope (the ‘Patlak time’) is an investigation independent function of scan time. This fact can be used to map SUR and SUV values from different investigations to a common time point for quantitative comparison. Additionally, it turns out that modelling the invariant AIF shape by an inverse power law is possible which further simplifies the correction procedure. The procedure was evaluated in 15 fully dynamic investigations of liver metastases from colorectal cancer and 10 dual time point (DTP) measurements. From each dynamic study, three ‘static scans’ at T=20,35,and 55 min post injection (p.i.) were created, where the last scan defined the reference time point to which the uptake values measured in the other two were corrected. The corrected uptake values were then compared to those actually measured at the reference time. For the DTP studies, the first scan (acquired at (78.1 ± 15.9) min p.i.) served as the reference, and the uptake values from the second scan (acquired (39.2 ± 9.9) min later) were corrected accordingly and compared to the reference.ResultsFor the dynamic data, the observed difference between uncorrected values and values at reference time was (-52±4.5)% at T=20 min and (-31±3.7)% at T=35 min for SUR and (-30±6.6)% at T=20 min and (-16±4)% at T=35 min for SUV. After correction, the difference was reduced to (-2.9±6.6)% at T=20 min and (-2.7±5)% at T=35 min for SUR and (1.9% ± 6.2)% at T=20 min and (1.7 ± 3.3)% at T=35 min for SUV. For the DTP studies, the observed differences of SUR and SUV between late and early scans were (48 ± 11)% and (24 ± 8.4)%, respectively. After correction, these differences were reduced to (2.6 ± 6.9)% and (-2.4±7.3)%, respectively.ConclusionIf FDG kinetics is irreversible in the targeted tissue, correction of SUV and SUR for scan time variability is possible with good accuracy. The correction distinctly improves comparability of lesion uptake values measured at different times post injection.

The influence of different signal-to-background ratios on spatial resolution and F18-FDG-PET quantification using point spread function and time-of-flight reconstruction

BackgroundF18-fluorodeoxyglucose positron-emission tomography (FDG-PET) reconstruction algorithms can have substantial influence on quantitative image data used, e.g., for therapy planning or monitoring in oncology. We analyzed radial activity concentration profiles of differently reconstructed FDG-PET images to determine the influence of varying signal-to-background ratios (SBRs) on the respective spatial resolution, activity concentration distribution, and quantification (standardized uptake value [SUV], metabolic tumor volume [MTV]).MethodsMeasurements were performed on a Siemens Biograph mCT 64 using a cylindrical phantom containing four spheres (diameter, 30 to 70 mm) filled with F18-FDG applying three SBRs (SBR1, 16:1; SBR2, 6:1; SBR3, 2:1). Images were reconstructed employing six algorithms (filtered backprojection [FBP], FBP + time-of-flight analysis [FBP + TOF], 3D-ordered subset expectation maximization [3D-OSEM], 3D-OSEM + TOF, point spread function [PSF], PSF + TOF). Spatial resolution was determined by fitting the convolution of the object geometry with a Gaussian point spread function to radial activity concentration profiles. MTV delineation was performed using fixed thresholds and semiautomatic background-adapted thresholding (ROVER, ABX, Radeberg, Germany).ResultsThe pairwise Wilcoxon test revealed significantly higher spatial resolutions for PSF + TOF (up to 4.0 mm) compared to PSF, FBP, FBP + TOF, 3D-OSEM, and 3D-OSEM + TOF at all SBRs (each P < 0.05) with the highest differences for SBR1 decreasing to the lowest for SBR3. Edge elevations in radial activity profiles (Gibbs artifacts) were highest for PSF and PSF + TOF declining with decreasing SBR (PSF + TOF largest sphere; SBR1, 6.3%; SBR3, 2.7%). These artifacts induce substantial SUVmax overestimation compared to the reference SUV for PSF algorithms at SBR1 and SBR2 leading to substantial MTV underestimation in threshold-based segmentation. In contrast, both PSF algorithms provided the lowest deviation of SUVmean from reference SUV at SBR1 and SBR2.ConclusionsAt high contrast, the PSF algorithms provided the highest spatial resolution and lowest SUVmean deviation from the reference SUV. In contrast, both algorithms showed the highest deviations in SUVmax and threshold-based MTV definition. At low contrast, all investigated reconstruction algorithms performed approximately equally. The use of PSF algorithms for quantitative PET data, e.g., for target volume definition or in serial PET studies, should be performed with caution - especially if comparing SUV of lesions with high and low contrasts.

Comparative evaluation of SUV, tumor-to-blood standard uptake ratio (SUR), and dual time point measurements for assessment of the metabolic uptake rate in FDG PET

BackgroundWe have demonstrated recently that the tumor-to-blood standard uptake ratio (SUR) is superior to tumor standardized uptake value (SUV) as a surrogate of the metabolic uptake rate Km of fluorodeoxyglucose (FDG), overcoming several of the known shortcomings of the SUV approach: excellent linear correlation of SUR and Km from Patlak analysis was found using dynamic imaging of liver metastases. However, due to the perfectly standardized uptake period used for SUR determination and the comparatively short uptake period, these results are not automatically valid and applicable for clinical whole-body examinations in which the uptake periods (T) are distinctly longer and can vary considerably. Therefore, the aim of this work was to investigate the correlation between SUR derived from clinical static whole-body scans and Km-surrogate derived from dual time point (DTP) measurements.MethodsDTP 18F-FDG PET/CT was performed in 90 consecutive patients with histologically proven non-small cell lung cancer (NSCLC). In the PET images, the primary tumor was delineated with an adaptive threshold method. For determination of the blood SUV, an aorta region of interest (ROI) was delineated manually in the attenuation CT and transferred to the PET image. Blood SUV was computed as the mean value of the aorta ROI. SUR values were computed as ratio of tumor SUV and blood SUV. SUR values from the early time point of each DTP measurement were scan time corrected to 75 min postinjection (SURtc). As surrogate of Km, we used the SUR(T) slope, Kslope, derived from DTP measurements since it is proportional to the latter under the given circumstances. The correlation of SUV and SURtc with Kslope was investigated. The prognostic value of SUV, SURtc, and Kslope for overall survival (OS) and progression-free survival (PFS) was investigated with univariate Cox regression in a homogeneous subgroup (N=31) treated with primary chemoradiation.ResultsCorrelation analysis revealed for both, SUV and SURtc, a clear linear correlation with Kslope (P<0.001). Correlation SUR vs. Kslope was considerably stronger than correlation SUV vs. Kslope (R2=0.92 and R2=0.69, respectively, P<0.001). Univariate Cox regression revealed SURtc and Kslope as significant prognostic factors for PFS (hazard ratio (HR) =3.4/ P=0.017 and HR =4.3/ P=0.020, respectively). For SUV, no significant effect was found. None of the investigated parameters was prognostic for OS.ConclusionsScan-time-corrected SUR is a significantly better surrogate of tumor FDG metabolism in clinical whole-body PET compared to SUV. The very high linear correlation of SUR and DTP-derived Kslope (which is proportional to actual Km) implies that for histologically proven malignant lesions, FDG-DTP does not provide added value in comparison to the SUR approach in NSCLC.

A volume of intersection approach for on-the-fly system matrix calculation in 3D PET image reconstruction.

The aim of this study is the evaluation of on-the-fly volume of intersection computation for systems geometry modelling in 3D PET image reconstruction. For this purpose we propose a simple geometrical model in which the cubic image voxels on the given Cartesian grid are approximated with spheres and the rectangular tubes of response (ToRs) are approximated with cylinders. The model was integrated into a fully 3D list-mode PET reconstruction for performance evaluation. In our model the volume of intersection between a voxel and the ToR is only a function of the impact parameter (the distance between voxel centre to ToR axis) but is independent of the relative orientation of voxel and ToR. This substantially reduces the computational complexity of the system matrix calculation. Based on phantom measurements it was determined that adjusting the diameters of the spherical voxel size and the ToR in such a way that the actual voxel and ToR volumes are conserved leads to the best compromise between high spatial resolution, low noise, and suppression of Gibbs artefacts in the reconstructed images. Phantom as well as clinical datasets from two different PET systems (Siemens ECAT HR(+) and Philips Ingenuity-TF PET/MR) were processed using the developed and the respective vendor-provided (line of intersection related) reconstruction algorithms. A comparison of the reconstructed images demonstrated very good performance of the new approach. The evaluation showed the respective vendor-provided reconstruction algorithms to possess 34-41% lower resolution compared to the developed one while exhibiting comparable noise levels. Contrary to explicit point spread function modelling our model has a simple straight-forward implementation and it should be easy to integrate into existing reconstruction software, making it competitive to other existing resolution recovery techniques.

Correction of quantification errors in pelvic and spinal lesions caused by ignoring higher photon attenuation of bone in [18F]NaF PET/MR.

PURPOSE MR-based attenuation correction (MRAC) in routine clinical whole-body positron emission tomography and magnetic resonance imaging (PET/MRI) is based on tissue type segmentation. Due to lack of MR signal in cortical bone and the varying signal of spongeous bone, standard whole-body segmentation-based MRAC ignores the higher attenuation of bone compared to the one of soft tissue (MRACnobone). The authors aim to quantify and reduce the bias introduced by MRACnobone in the standard uptake value (SUV) of spinal and pelvic lesions in 20 PET/MRI examinations with [18F]NaF. METHODS The authors reconstructed 20 PET/MR [18F]NaF patient data sets acquired with a Philips Ingenuity TF PET/MRI. The PET raw data were reconstructed with two different attenuation images. First, the authors used the vendor-provided MRAC algorithm that ignores the higher attenuation of bone to reconstruct PETnobone. Second, the authors used a threshold-based algorithm developed in their group to automatically segment bone structures in the [18F]NaF PET images. Subsequently, an attenuation coefficient of 0.11 cm(-1) was assigned to the segmented bone regions in the MRI-based attenuation image (MRACbone) which was used to reconstruct PETbone. The automatic bone segmentation algorithm was validated in six PET/CT [18F]NaF examinations. Relative SUVmean and SUVmax differences between PETbone and PETnobone of 8 pelvic and 41 spinal lesions, and of other regions such as lung, liver, and bladder, were calculated. By varying the assigned bone attenuation coefficient from 0.11 to 0.13 cm(-1), the authors investigated its influence on the reconstructed SUVs of the lesions. RESULTS The comparison of [18F]NaF-based and CT-based bone segmentation in the six PET/CT patients showed a Dice similarity of 0.7 with a true positive rate of 0.72 and a false discovery rate of 0.33. The [18F]NaF-based bone segmentation worked well in the pelvis and spine. However, it showed artifacts in the skull and in the extremities. The analysis of the 20 [18F]NaF PET/MRI examinations revealed relative SUVmax differences between PETnobone and PETbone of (-8.8%±2.7%, p=0.01) and (-8.1%±1.9%, p=2.4×10(-8)) in pelvic and spinal lesions, respectively. A maximum SUVmax underestimation of -13.7% was found in lesion in the third cervical spine. The averaged SUVmean differences in volumes of interests in lung, liver, and bladder were below 3%. The average SUVmax differences in pelvic and spinal lesions increased from -9% to -18% and -8% to -17%, respectively, when increasing the assigned bone attenuation coefficient from 0.11 to 0.13 cm(-1). CONCLUSIONS The developed automatic [18F]NaF PET-based bone segmentation allows to include higher bone attenuation in whole-body MRAC and thus improves quantification accuracy for pelvic and spinal lesions in [18F]NaF PET/MRI examinations. In nonbone structures (e.g., lung, liver, and bladder), MRACnobone yields clinically acceptable accuracy.

Preliminary evaluation of the MLAA algorithm with the Philips Ingenuity PET/MR.

Combined PET/MR is a promising tool for simultaneous investigation of soft tissue morphology and function. However, contrary to CT, MR images do not provide information on photon attenuation in tissue. In the currently available systems issue is solved by synthesizing attenuation maps from MR images using segmentation algorithms. This approach has been shown to provide reason-able results in most cases. However, sporadically occurring segmentation errors can cause serious problems. Recently, algorithms for simultaneous estimation of attenuation and tracer distribution (MLAA) have been introduced. So far, validity of MLAA has mainly been demonstrated in simulated data. We have integrated the MLAA algorithm [2] into the THOR reconstruction [1]. An evaluation of MLAA was performed using both phantom and patient data acquired with the Ingenuity PET/MR. Phantom data were acquired using a whole body phantom with three cylindrical inserts filled with different substances (plastic, air, glycerol). MLAA-estimated mu-maps of the phantom were compared to the mu-maps resulting from transmission measurements with an ECAT HR+ scanner. We also performed a first qualitative evaluation of the attenuation maps obtained in patient studies. Evaluation of the phantom study showed good concordance between measured and estimated attenuation coefficients for all types of substances used in the phantom. Evaluation of patient data showed some substantial improvements of the MLAA attenuation maps compared to the segmented MR-based attenuation maps. Preliminary results show that for the Philips Ingenuity PET/MR scanner the MLAA algorithm allows to obtain attenuation maps which outperform the MR based maps in several aspects. However, a more detailed analysis is still required to address the question of possible cross-talks in regions with high activity. Additionally, MLAA algorithm substantially increases computational burden leading to long processing times, which makes it currently impractical for clinical application.

Comparison of different tube-of-response (TOR) models for resolution recovery in PET image reconstruction for the Philips Ingenuity TF PET/MR

Recently, we have proposed a method for on-the-fly system matrix computation where the tube-of-response (TOR) is approximated as a cylinder with constant density (TORCD) and the cubic voxels are replaced by spheres. We could show that with this model the PET image quality can be notably improved compared to the vendor provided image reconstruction of our Philips Ingenuity-TF PET/MR. In this work we address the question whether image quality can be further improved by using a variable density TOR (TOR-VD). The radial variability of TOR-VD was modelled by a Kaiser-Bessel function. Free parameters of this density model were used to optimize image properties regarding resolution, noise, and Gibbs artifacts. Additional, a TOR-VD model accounting for position dependent effects along the TOR caused by the finite solid angles of the detectors is under investigation. Phantom measurement were performed with a Philips Ingenuity-TF PET/MR scanner. Listmode data were reconstructed using TOR-CD and TORVD, respectively on two different grids with cubic voxel size of 2 mm and 4 mm. Image quality was assessed with resolution-noise curves and investigation of the radial position dependence of the spatial resolution. For 2 mm voxels, TOR-VD consistently yields a slight improvement of the investigated image quality measures compared to TOR-CD. For 4 mm voxels both models lead essentially to the same results. These findings can be understood as a consequence of the relative size of voxel and TOR. For typical whole body studies (4 mm voxel size) a variable TOR does not improve image quality beyond what is achievable with a constant density TOR. For smaller voxel size the image quality can indeed be somewhat improved with a variable TOR but at the expense of drastically increased computation time.