C. Marshall

ATLAAS: an automatic decision tree-based learning algorithm for advanced image segmentation in positron emission tomography.

Accurate and reliable tumour delineation on positron emission tomography (PET) is crucial for radiotherapy treatment planning. PET automatic segmentation (PET-AS) eliminates intra- and interobserver variability, but there is currently no consensus on the optimal method to use, as different algorithms appear to perform better for different types of tumours. This work aimed to develop a predictive segmentation model, trained to automatically select and apply the best PET-AS method, according to the tumour characteristics. ATLAAS, the automatic decision tree-based learning algorithm for advanced segmentation is based on supervised machine learning using decision trees. The model includes nine PET-AS methods and was trained on a 100 PET scans with known true contour. A decision tree was built for each PET-AS algorithm to predict its accuracy, quantified using the Dice similarity coefficient (DSC), according to the tumour volume, tumour peak to background SUV ratio and a regional texture metric. The performance of ATLAAS was evaluated for 85 PET scans obtained from fillable and printed subresolution sandwich phantoms. ATLAAS showed excellent accuracy across a wide range of phantom data and predicted the best or near-best segmentation algorithm in 93% of cases. ATLAAS outperformed all single PET-AS methods on fillable phantom data with a DSC of 0.881, while the DSC for H&N phantom data was 0.819. DSCs higher than 0.650 were achieved in all cases. ATLAAS is an advanced automatic image segmentation algorithm based on decision tree predictive modelling, which can be trained on images with known true contour, to predict the best PET-AS method when the true contour is unknown. ATLAAS provides robust and accurate image segmentation with potential applications to radiation oncology.

Development and validation of a prognostic model incorporating texture analysis derived from standardised segmentation of pet in patients with oesophageal cancer

ObjectivesThis retrospective cohort study developed a prognostic model incorporating PET texture analysis in patients with oesophageal cancer (OC). Internal validation of the model was performed.MethodsConsecutive OC patients (n = 403) were chronologically separated into development (n = 302, September 2010-September 2014, median age = 67.0, males = 227, adenocarcinomas = 237) and validation cohorts (n = 101, September 2014-July 2015, median age = 69.0, males = 78, adenocarcinomas = 79). Texture metrics were obtained using a machine-learning algorithm for automatic PET segmentation. A Cox regression model including age, radiological stage, treatment and 16 texture metrics was developed. Patients were stratified into quartiles according to a prognostic score derived from the model. A p-value < 0.05 was considered statistically significant. Primary outcome was overall survival (OS).ResultsSix variables were significantly and independently associated with OS: age [HR =1.02 (95% CI 1.01-1.04), p < 0.001], radiological stage [1.49 (1.20-1.84), p < 0.001], treatment [0.34 (0.24–0.47), p < 0.001], log(TLG) [5.74 (1.44–22.83), p = 0.013], log(Histogram Energy) [0.27 (0.10–0.74), p = 0.011] and Histogram Kurtosis [1.22 (1.04–1.44), p = 0.017]. The prognostic score demonstrated significant differences in OS between quartiles in both the development (X2 143.14, df 3, p < 0.001) and validation cohorts (X2 20.621, df 3, p < 0.001).ConclusionsThis prognostic model can risk stratify patients and demonstrates the additional benefit of PET texture analysis in OC staging.Key points• PET texture analysis adds prognostic value to oesophageal cancer staging.• Texture metrics are independently and significantly associated with overall survival.• A prognostic model including texture analysis can help risk stratify patients.

Head and neck target delineation using a novel PET automatic segmentation algorithm

PURPOSE To evaluate the feasibility and impact of using a novel advanced PET auto-segmentation method in Head and Neck (H&N) radiotherapy treatment (RT) planning. METHODS ATLAAS, Automatic decision Tree-based Learning Algorithm for Advanced Segmentation, previously developed and validated on pre-clinical data, was applied to 18F-FDG-PET/CT scans of 20 H&N patients undergoing Intensity Modulated Radiation Therapy. Primary Gross Tumour Volumes (GTVs) manually delineated on CT/MRI scans (GTVpCT/MRI), together with ATLAAS-generated contours (GTVpATLAAS) were used to derive the RT planning GTV (GTVpfinal). ATLAAS outlines were compared to CT/MRI and final GTVs qualitatively and quantitatively using a conformity metric. RESULTS The ATLAAS contours were found to be reliable and useful. The volume of GTVpATLAAS was smaller than GTVpCT/MRI in 70% of the cases, with an average conformity index of 0.70. The information provided by ATLAAS was used to grow the GTVpCT/MRI in 10 cases (up to 10.6mL) and to shrink the GTVpCT/MRI in 7 cases (up to 12.3mL). ATLAAS provided complementary information to CT/MRI and GTVpATLAAS contributed to up to 33% of the final GTV volume across the patient cohort. CONCLUSIONS ATLAAS can deliver operator independent PET segmentation to augment clinical outlining using CT and MRI and could have utility in future clinical studies.

A novel phantom technique for evaluating the performance of PET auto-segmentation methods in delineating heterogeneous and irregular lesions.

BackgroundPositron Emission Tomography (PET)-based automatic segmentation (PET-AS) methods can improve tumour delineation for radiotherapy treatment planning, particularly for Head and Neck (H&N) cancer. Thorough validation of PET-AS on relevant data is currently needed. Printed subresolution sandwich (SS) phantoms allow modelling heterogeneous and irregular tracer uptake, while providing reference uptake data. This work aimed to demonstrate the usefulness of the printed SS phantom technique in recreating complex realistic H&N radiotracer uptake for evaluating several PET-AS methods.MethodsTen SS phantoms were built from printouts representing 2mm-spaced slices of modelled H&N uptake, printed using black ink mixed with 18F-fluorodeoxyglucose, and stacked between 2mm thick plastic sheets. Spherical lesions were modelled for two contrasted uptake levels, and irregular and spheroidal tumours were modelled for homogeneous, and heterogeneous uptake including necrotic patterns. The PET scans acquired were segmented with ten custom PET-AS methods: adaptive iterative thresholding (AT), region growing, clustering applied to 2 to 8 clusters, and watershed transform-based segmentation. The difference between the resulting contours and the ground truth from the image template was evaluated using the Dice Similarity Coefficient (DSC), Sensitivity and Positive Predictive value.ResultsRealistic H&N images were obtained within 90 min of preparation. The sensitivity of binary PET-AS and clustering using small numbers of clusters dropped for highly heterogeneous spheres. The accuracy of PET-AS methods dropped between 4% and 68% for irregular lesions compared to spheres of the same volume. For each geometry and uptake modelled with the SS phantoms, we report the number of clusters resulting in optimal segmentation. Radioisotope distributions representing necrotic uptakes proved most challenging for most methods. Two PET-AS methods did not include the necrotic region in the segmented volume.ConclusionsPrinted SS phantoms allowed identifying advantages and drawbacks of the different methods, determining the most robust PET-AS for the segmentation of heterogeneities and complex geometries, and quantifying differences across methods in the delineation of necrotic lesions. The printed SS phantom technique provides key advantages in the development and evaluation of PET segmentation methods and has a future in the field of radioisotope imaging.

Evaluation of prognostic models developed using standardised image features from different PET automated segmentation methods

BackgroundPrognosis in oesophageal cancer (OC) is poor. The 5-year overall survival (OS) rate is approximately 15%. Personalised medicine is hoped to increase the 5- and 10-year OS rates. Quantitative analysis of PET is gaining substantial interest in prognostic research but requires the accurate definition of the metabolic tumour volume. This study compares prognostic models developed in the same patient cohort using individual PET segmentation algorithms and assesses the impact on patient risk stratification.Consecutive patients (n = 427) with biopsy-proven OC were included in final analysis. All patients were staged with PET/CT between September 2010 and July 2016. Nine automatic PET segmentation methods were studied. All tumour contours were subjectively analysed for accuracy, and segmentation methods with < 90% accuracy were excluded. Standardised image features were calculated, and a series of prognostic models were developed using identical clinical data. The proportion of patients changing risk classification group were calculated.ResultsOut of nine PET segmentation methods studied, clustering means (KM2), general clustering means (GCM3), adaptive thresholding (AT) and watershed thresholding (WT) methods were included for analysis. Known clinical prognostic factors (age, treatment and staging) were significant in all of the developed prognostic models. AT and KM2 segmentation methods developed identical prognostic models. Patient risk stratification was dependent on the segmentation method used to develop the prognostic model with up to 73 patients (17.1%) changing risk stratification group.ConclusionPrognostic models incorporating quantitative image features are dependent on the method used to delineate the primary tumour. This has a subsequent effect on risk stratification, with patients changing groups depending on the image segmentation method used.

Optimising graphical techniques applied to irreversible tracers

Graphical analysis techniques are often applied to positron emission tomography (PET) images to estimate physiological parameters. Patlak analysis is primarily used to obtain the rate constant (Ki) that indicates the transfer of a tracer from plasma to the irreversible compartment and ultimately describes how the tracer binds to the targeted tissue. One of the most common issues associated with Patlak analysis is the introduction of statistical noise that affects the slope of the graphical plot and causes bias. In this study, several statistical methods are proposed and applied to PET time activity curves (TACs) for both reversible and irreversible regions that are involved in the equation. A dynamic PET imaging simulator for the Patlak model was used to evaluate the statistical methods employed to reduce the bias introduced in the acquired data.

Optimization of Zirconium-89 production in IBA cyclone 18/9 cyclotron with COSTIS solid target system

Zirconium-89 is a promising radionuclide in the development of new immuno-PET agents for in vivo imaging of cancerous tumours and radioimmunotherapy (RIT) planning. Besides the convenient half-life of 78.4 h, 89Zr has a beta plus emission rate of 23% and a low maximum energy of 0.9 MeV, delivering good spatial resolution as a result of short positron range in tissue (around 1 mm). Cyclotron production for the radiometal of 89Zr was investigated to find optimal conditions according to results of FLUKA code Monte Carlo modelling of irradiation processes, nuclear reactions and target design. This was followed by reasonably detailed experimental validation (making cyclotron productions for expected high product yield and low impurities levels followed by activity measurements, spectra acquisitions and chemical separation procedures), in which the strategies developed by computer models were carried out in the IBA Cyclone 18/9 cyclotron, permitting a comparison of the predicted and actual yields of 89Zr and isotopic by-products (impurities). Once the in silica model was validated experimentally, then optimal method of the radiometal production in the cyclotron was developed.

EP-1333: Impact of 18 F-Choline PET scan acquisition time on delineation of GTV in Prostate cancer

Background: Dose painting radiotherapy requires accurate outlining of primary tumour volumes in the prostate. T2-Weighted (T2W) Magnetic Resonance Imaging (MRI) is the best imaging method for defining the gross tumour volume (GTV). Choline positron emission tomography (PET) is currently a controversial tracer. The image acquisition differs significantly in published studies. Many used early static imaging. One study found that 18F-choline PET/CT with late image acquisition has superior accuracy to T2W MR and functional MR alone1. We investigate whether increasing 18F-Choline PET scan acquisition time from 60 (PET-60) to 90 (PET-90) minutes improves GTV TVD. Methods. Analysis was performed on 9 18F-Choline PET scans. Patients were injected with 370MBq of activity. Three clinicians (C1, C2 and C3) independently and without reference to each other contoured GTVs on each of the T2W-MRI, PET-60 and PET-90 scans at differing times. Scans were registered by a clinician using rigid co-registration. The treating clinicians MRI contour was used as a reference contour. The resulting PET and MRI GTVs were transferred to the PET-60 and PET-90 scans after image registration. The Dice Similarity Coefficient (DSC), Specificity (Sp) and Sensitivity (S) were calculated from contour mask voxel analysis. Results. Table 1 shows the mean and range DSC, S and Sp scores on MRI, PET-60 and PET-90 for C1, C2 and C3 in comparison to the treating clinicians contour on MRI (C1). A 2 sampled T-test (P < 0.01) showed, no significant difference in the Sp, S and DSC between GTVs on PET-60 and PET-90 scans. Further to this, as shown in Figure 1, variability in GTV delineation is significant between observers in a singular case as well as across imaging modalities. Conclusion. Compared to MRI delineated GTVs, 18F-Choline PET GTVs are significantly different. This study found however that increasing the PET scan acquisition time from 60 to 90 minutes did not improve the performance of GTV TVD in comparison to MRI delineated GTV.

ESTABLISHMENT OF A PET RADIOTRACER NETWORK FOR DEMENTIA RESEARCH

differences in ADAS-cog sub-scores between the typical (Figure 1A) and parietal (Figure ID) subgroups estimated by SuStaln (p<0.05 shaded in blue). Parietal 1⁄4 AD patients with a higher probability of belonging to the parietal subtype (Figure 1D) than any of the other subtypes; Typical 1⁄4 AD patients with a higher probability of belonging to the typical subtype (Figure 1A) than any of the other subtypes; Strong Parietal 1⁄4 AD patients with probability>0.75 of belonging to the parietal subtype (Figure 1D); Strong Typical 1⁄4 AD patients with probability>0.75 of belonging to the typical subtype (Figure 1A). Poster Presentations: Saturday, July 15, 2017 P117

EP-1872: Benchmarking texture analysis for PET in oesophageal cancer

Material and Methods: Twelve consecutive patients with a histopathological diagnosis of stage I-IV MPM (6 left-sided and 6 right-sided) were included. CT scans with IV contrast, 18FFDG PET/CT scans and MRI scans (post-contrast T1-weighted, T2 and diffusion-weighted images [DWI]) were obtained and downloaded from the institutional database onto a standalone image fusion workstation (MIM Software Inc., Cleveland, OH, USA) for image registration and contouring. CT scans were rigidly co-registered with 18FDG-CT-PET, with MRI scans and with DWI scans. Four sets of pleural GTVs were defined: 1) a CT-based GTV (GTVCT); 2)a PET/CT-based GTV (GTVCT+PET/CT); 3) a T1/T2-weighted MRI-based GTV (GTVCT+MRI); 4) a DWI-based GTV (GTVCT+DWI). Only the pleural tumor was contoured; mediastinal nodes were excluded. “Quantitative” and “qualitative” (visual) evaluation of the volumes was performed.