Search

Artificial intelligence (AI)-based methods are showing substantial promise in segmenting oncologic positron emission tomography (PET) images. For clinical translation of these methods, assessing their performance on clinically relevant tasks is important. However, these methods are typically evaluated using metrics that may not correlate with the task performance. One such widely used metric is the Dice score, a figure of merit that measures the spatial overlap between the estimated segmentation and a reference standard (e.g., manual segmentation). In this work, we investigated whether evaluating AI-based segmentation methods using Dice scores yields a similar interpretation as evaluation on the clinical tasks of quantifying metabolic tumor volume (MTV) and total lesion glycolysis (TLG) of primary tumor from PET images of patients with non-small cell lung cancer. The investigation was conducted via a retrospective analysis with the ECOG-ACRIN 6668/RTOG 0235 multi-center clinical trial data. Specifically, we evaluated different structures of a commonly used AI-based segmentation method using both Dice scores and the accuracy in quantifying MTV/TLG. Our results show that evaluation using Dice scores can lead to findings that are inconsistent with evaluation using the task-based figure of merit. Thus, our study motivates the need for objective task-based evaluation of AI-based segmentation methods for quantitative PET.

Objective evaluation of quantitative imaging (QI) methods with patient data is highly desirable, but is hindered by the lack or unreliability of an available gold standard. To address this issue, techniques that can evaluate QI methods without access to a gold standard are being actively developed. These techniques assume that the true and measured values are linearly related by a slope, bias, and Gaussian-distributed noise term, where the noise between measurements made by different methods is independent of each other. However, this noise arises in the process of measuring the same quantitative value, and thus can be correlated. To address this limitation, we propose a no-gold-standard evaluation (NGSE) technique that models this correlated noise by a multi-variate Gaussian distribution parameterized by a covariance matrix. We derive a maximum-likelihood-based approach to estimate the parameters that describe the relationship between the true and measured values, without any knowledge of the true values. We then use the estimated slopes and diagonal elements of the covariance matrix to compute the noise-to-slope ratio (NSR) to rank the QI methods on the basis of precision. The proposed NGSE technique was evaluated with multiple numerical experiments. Our results showed that the technique reliably estimated the NSR values and yielded accurate rankings of the considered methods for ~ 83% of 160 trials. In particular, the technique correctly identified the most precise method for ~ 97% of the trials. Overall, this study demonstrates the efficacy of the NGSE technique to accurately rank different QI methods when the correlated noise is present, and without access to any knowledge of the ground truth. The results motivate further validation of this technique with realistic simulation studies and patient data.

Artificial intelligence (AI)-based methods are showing promise in multiple medical-imaging applications. Thus, there is substantial interest in clinical translation of these methods, requiring in turn, that they be evaluated rigorously. In this paper, our goal is to lay out a framework for objective task-based evaluation of AI methods. We will also provide a list of tools available in the literature to conduct this evaluation. Further, we outline the important role of physicians in conducting these evaluation studies. The examples in this paper will be proposed in the context of PET with a focus on neural-network-based methods. However, the framework is also applicable to evaluate other medical-imaging modalities and other types of AI methods.

Objective evaluation of new and improved methods for PET imaging requires access to images with ground truth, as can be obtained through simulation studies. However, for these studies to be clinically relevant, it is important that the simulated images are clinically realistic. In this study, we develop a stochastic and physics-based method to generate realistic oncological two-dimensional (2-D) PET images, where the ground-truth tumor properties are known. The developed method extends upon a previously proposed approach. The approach captures the observed variabilities in tumor properties from actual patient population. Further, we extend that approach to model intra-tumor heterogeneity using a lumpy object model. To quantitatively evaluate the clinical realism of the simulated images, we conducted a human-observer study. This was a two-alternative forced-choice (2AFC) study with trained readers (five PET physicians and one PET physicist). Our results showed that the readers had an average of ~ 50% accuracy in the 2AFC study. Further, the developed simulation method was able to generate wide varieties of clinically observed tumor types. These results provide evidence for the application of this method to 2-D PET imaging applications, and motivate development of this method to generate 3-D PET images.

Quantitative measures of dopamine transporter (DaT) uptake in caudate, putamen, and globus pallidus (GP) have potential as biomarkers for measuring the severity of Parkinson disease. Reliable quantification of this uptake requires accurate segmentation of the considered regions. However, segmentation of these regions from DaT-SPECT images is challenging, a major reason being partial-volume effects (PVEs), which arise from the limited system resolution and reconstruction of images over finite-sized voxel grids. The latter leads to tissue-fraction effects (TFEs). Thus, there is an important need for methods that can account for the PVEs, including the TFEs, and accurately segment DaT-SPECT images. The purpose of this study is to design and objectively evaluate a fully automated tissue-fraction estimation-based segmentation method that segments the caudate, putamen, and GP from DaT-SPECT images. The proposed method estimates the posterior mean of the fractional volumes occupied by the caudate, putamen, and GP within each voxel of a 3-D DaT-SPECT image. The estimate is obtained by minimizing a cost function based on the binary cross-entropy loss between the true and estimated fractional volumes over a population of SPECT images. Evaluations using clinically guided highly realistic simulation studies show that the proposed method accurately segmented the caudate, putamen, and GP with high mean Dice similarity coefficients ~ 0.80 and significantly outperformed (p < 0.01) all other considered segmentation methods. Further, objective evaluation of the proposed method on the task of quantifying regional uptake shows that the method yielded reliable quantification with low ensemble normalized root mean square error (NRMSE) < 20% for all the considered regions. The results motivate further evaluation of the method with physical-phantom and patient studies.

Objective evaluation of quantitative imaging (QI) methods using measurements directly obtained from patient images is highly desirable but hindered by the non-availability of gold standards. To address this issue, statistical techniques have been proposed to objectively evaluate QI methods without a gold standard. These techniques assume that the measured and true values are linearly related by a slope, bias, and normally distributed noise term, where it is assumed that the noise term between the different methods is independent. However, the noise could be correlated since it arises in the process of measuring the same true value. To address this issue, we propose a new no-gold-standard evaluation (NGSE) technique that models this noise as a multivariate normally distributed term, characterized by a covariance matrix. In this manuscript, we derive a maximum-likelihood-based technique that, without any knowledge of the true QI values, estimates the slope, bias, and covariance matrix terms. These are then used to rank the methods on the basis of precision of the measured QI values. Overall, the derivation demonstrates the mathematical premise behind the proposed NGSE technique.

Tumor segmentation in oncological PET is challenging, a major reason being the partial-volume effects that arise due to low system resolution and finite voxel size. The latter results in tissue-fraction effects, i.e. voxels contain a mixture of tissue classes. Conventional segmentation methods are typically designed to assign each voxel in the image as belonging to a certain tissue class. Thus, these methods are inherently limited in modeling tissue-fraction effects. To address the challenge of accounting for partial-volume effects, and in particular, tissue-fraction effects, we propose a Bayesian approach to tissue-fraction estimation for oncological PET segmentation. Specifically, this Bayesian approach estimates the posterior mean of fractional volume that the tumor occupies within each voxel of the image. The proposed method, implemented using a deep-learning-based technique, was first evaluated using clinically realistic 2-D simulation studies with known ground truth, in the context of segmenting the primary tumor in PET images of patients with lung cancer. The evaluation studies demonstrated that the method accurately estimated the tumor-fraction areas and significantly outperformed widely used conventional PET segmentation methods, including a U-net-based method, on the task of segmenting the tumor. In addition, the proposed method was relatively insensitive to partial-volume effects and yielded reliable tumor segmentation for different clinical-scanner configurations. The method was then evaluated using clinical images of patients with stage IIB/III non-small cell lung cancer from ACRIN 6668/RTOG 0235 multi-center clinical trial. Here, the results showed that the proposed method significantly outperformed all other considered methods and yielded accurate tumor segmentation on patient images with Dice similarity coefficient (DSC) of 0.82 (95 % CI: [0.78, 0.86]).