Finding a noninvasive radiomic surrogate of tumor immune features could help identify patients more likely to respond to novel immune checkpoint inhibitors. Particularly, CD73 is an ectonucleotidase ...that catalyzes the breakdown of extracellular AMP into immunosuppressive adenosine, which can be blocked by therapeutic antibodies. High CD73 expression in colorectal cancer liver metastasis (CRLM) resected with curative intent is associated with early recurrence and shorter patient survival. The aim of this study was hence to evaluate whether machine learning analysis of preoperative liver CT-scan could estimate high vs low CD73 expression in CRLM and whether such radiomic score would have a prognostic significance.
We trained an Attentive Interpretable Tabular Learning (TabNet) model to predict, from preoperative CT images, stratified expression levels of CD73 (CD73
vs. CD73
) assessed by immunofluorescence (IF) on tissue microarrays. Radiomic features were extracted from 160 segmented CRLM of 122 patients with matched IF data, preprocessed and used to train the predictive model. We applied a five-fold cross-validation and validated the performance on a hold-out test set.
TabNet provided areas under the receiver operating characteristic curve of 0.95 (95% CI 0.87 to 1.0) and 0.79 (0.65 to 0.92) on the training and hold-out test sets respectively, and outperformed other machine learning models. The TabNet-derived score, termed rad-CD73, was positively correlated with CD73 histological expression in matched CRLM (Spearman's ρ = 0.6004; P < 0.0001). The median time to recurrence (TTR) and disease-specific survival (DSS) after CRLM resection in rad-CD73
vs rad-CD73
patients was 13.0 vs 23.6 months (P = 0.0098) and 53.4 vs 126.0 months (P = 0.0222), respectively. The prognostic value of rad-CD73 was independent of the standard clinical risk score, for both TTR (HR = 2.11, 95% CI 1.30 to 3.45, P < 0.005) and DSS (HR = 1.88, 95% CI 1.11 to 3.18, P = 0.020).
Our findings reveal promising results for non-invasive CT-scan-based prediction of CD73 expression in CRLM and warrant further validation as to whether rad-CD73 could assist oncologists as a biomarker of prognosis and response to immunotherapies targeting the adenosine pathway.
Purpose:
Accurate geometry is required for radiotherapy treatment planning (RTP). When considering the use of magnetic resonance imaging (MRI) for RTP, geometric distortions observed in the acquired ...images should be considered. While scanner technology and vendor supplied correction algorithms provide some correction, large distortions are still present in images, even when considering considerably smaller scan lengths than those typically acquired with CT in conventional RTP. This study investigates MRI acquisition with a moving table compared with static scans for potential geometric benefits for RTP.
Methods:
A full field of view (FOV) phantom (diameter 500 mm; length 513 mm) was developed for measuring geometric distortions in MR images over volumes pertinent to RTP. The phantom consisted of layers of refined plastic within which vitamin E capsules were inserted. The phantom was scanned on CT to provide the geometric gold standard and on MRI, with differences in capsule location determining the distortion. MRI images were acquired with two techniques. For the first method, standard static table acquisitions were considered. Both 2D and 3D acquisition techniques were investigated. With the second technique, images were acquired with a moving table. The same sequence was acquired with a static table and then with table speeds of 1.1 mm/s and 2 mm/s. All of the MR images acquired were registered to the CT dataset using a deformable B‐spline registration with the resulting deformation fields providing the distortion information for each acquisition.
Results:
MR images acquired with the moving table enabled imaging of the whole phantom length while images acquired with a static table were only able to image 50%–70% of the phantom length of 513 mm. Maximum distortion values were reduced across a larger volume when imaging with a moving table. Increased table speed resulted in a larger contribution of distortion from gradient nonlinearities in the through‐plane direction and an increased blurring of capsule images, resulting in an apparent capsule volume increase by up to 170% in extreme axial FOV regions. Blurring increased with table speed and in the central regions of the phantom, geometric distortion was less for static table acquisitions compared to a table speed of 2 mm/s over the same volume. Overall, the best geometric accuracy was achieved with a table speed of 1.1 mm/s.
Conclusions:
The phantom designed enables full FOV imaging for distortion assessment for the purposes of RTP. MRI acquisition with a moving table extends the imaging volume in the z direction with reduced distortions which could be useful particularly if considering MR‐only planning. If utilizing MR images to provide additional soft tissue information to the planning CT, standard acquisition sequences over a smaller volume would avoid introducing additional blurring or distortions from the through‐plane table movement.
Immune checkpoint blockade has not yet been effective in patients with mismatch repair proficient metastatic colorectal cancer. Targeting immunosuppressive metabolic pathways is being explored as a ...new immunotherapeutic approach. We assessed whether CD73, the rate limiting enzyme that catalyzes the degradation of extracellular AMP into immunosuppressive adenosine, could be an immunological determinant of colorectal liver metastases (CRLMs). By immunofluorescence on tissue microarrays, intratumoral CD73 expression (tCD73) was analyzed in 391 CRLMs resected in 215 patients, and soluble CD73 (sCD73) was measured by ELISA in the pre-operative serum of 193 patients. High tCD73 was associated with worse pathological features, such as multiple and larger CRLMs, and poorer pathologic response to pre-operative chemotherapy. The median time to recurrence and disease-specific survival after CRLM resection was significantly shorter in patients with high tCD73 (11.0 and 46.4 months, respectively) compared with low tCD73 (19.0 and 61.5 months, respectively). tCD73 was strongly associated with patient outcomes independently of clinicopathological variables. sCD73 did not correlate with tCD73. Patients with high levels of sCD73 also had shorter disease-specific survival. Our results suggested that CD73 in CRLMs may be prognostically informative and may help select patients more likely to respond to adenosine pathway blocking agents.
In colorectal cancer liver metastases (CRLM), the density of tumor-infiltrating lymphocytes, the expression of class I major histocompatibility complex (MHC-I), and the pathological response to ...preoperative chemotherapy have been associated with oncological outcomes after complete resection. However, the prognostic significance of the heterogeneity of these features in patients with multiple CRLMs remains under investigation. We used a tissue microarray of 220 mismatch repair-gene proficient CRLMs resected in 97 patients followed prospectively to quantify CD3
+
T cells and MHC-I by immunohistochemistry. Histopathological response to preoperative chemotherapy was assessed using standard scoring systems. We tested associations between clinical, immunological, and pathological features with oncologic outcomes. Overall, 29 patients (30.2%) had CRLMs homogeneous for CD3+ T cell infiltration and MHC-I. Patients with immune homogeneous compared to heterogeneous CRLMs had longer median time to recurrence (TTR) (30 vs. 12 months,
p
= .0018) and disease-specific survival (DSS) (not reached vs. 48 months,
p
= .0009). At 6 years, 80% of the patients with immune homogeneous CRLMs were still alive. Homogeneity of response to preoperative chemotherapy was seen in 60 (61.9%) and 69 (80.2%) patients according to different grading systems and was not associated with TTR or DSS. CD3 and MHC-I heterogeneity was independent of response to pre-operative chemotherapy and of other clinicopathological variables for their association with oncological outcomes. In patients with multiple CRLMs resected with curative intent, similar adaptive immune features seen across metastases could be more informative than pathological response to pre-operative chemotherapy in predicting oncological outcomes.
The purpose of this study was to determine the impact of magnetic resonance imaging (MRI) geometric distortions when using MRI for target delineation and planning for whole‐breast, ...intensity‐modulated radiotherapy (IMRT). Residual system distortions and combined systematic and patient‐induced distortions are considered. This retrospective study investigated 18 patients who underwent whole‐breast external beam radiotherapy, where both CT and MRIs were acquired for treatment planning. Distortion phantoms were imaged on two MRI systems, dedicated to radiotherapy planning (a wide, closed‐bore 3T and an open‐bore 1T). Patient scans were acquired on the 3T system. To simulate MRI‐based planning, distortion maps representing residual system distortions were generated via deformable registration between phantom CT and MRIs. Patient CT images and structures were altered to match the residual system distortion measured by the phantoms on each scanner. The patient CTs were also registered to the corresponding patient MRI scans, to assess patient and residual system effects. Tangential IMRT plans were generated and optimized on each resulting CT dataset, then propagated to the original patient CT space. The resulting dose distributions were then evaluated with respect to the standard clinically acceptable DVH and visual assessment criteria. Maximum residual systematic distortion was measured to be 7.9 mm (95%<4.7mm) and 11.9 mm (95%<4.6mm) for the 3T and 1T scanners, respectively, which did not result in clinically unacceptable plans. Eight of the plans accounting for patient and systematic distortions were deemed clinically unacceptable when assessed on the original CT. For these plans, the mean difference in PTV V95 (volume receiving 95% prescription dose) was 0.13±2.51% and −0.73±1.93% for right‐ and left‐sided patients, respectively. Residual system distortions alone had minimal impact on the dosimetry for the two scanners investigated. The combination of MRI systematic and patient‐related distortions can result in unacceptable dosimetry for whole‐breast IMRT, a potential issue when considering MRI‐only radiotherapy treatment planning.
PACS number(s): 87.61.‐c, 87.57.cp, 87.57.nj, 87.55.D‐
To validate automatic substitute computed tomography CT (sCT) scans generated from standard T2-weighted (T2w) magnetic resonance (MR) pelvic scans for MR-Sim prostate treatment planning.
A Siemens ...Skyra 3T MR imaging (MRI) scanner with laser bridge, flat couch, and pelvic coil mounts was used to scan 39 patients scheduled for external beam radiation therapy for localized prostate cancer. For sCT generation a whole-pelvis MRI scan (1.6 mm 3-dimensional isotropic T2w SPACE Sampling Perfection with Application optimized Contrasts using different flip angle Evolution sequence) was acquired. Three additional small field of view scans were acquired: T2w, T2*w, and T1w flip angle 80° for gold fiducials. Patients received a routine planning CT scan. Manual contouring of the prostate, rectum, bladder, and bones was performed independently on the CT and MR scans. Three experienced observers contoured each organ on MRI, allowing interobserver quantification. To generate a training database, each patient CT scan was coregistered to their whole-pelvis T2w using symmetric rigid registration and structure-guided deformable registration. A new multi-atlas local weighted voting method was used to generate automatic contours and sCT results.
The mean error in Hounsfield units between the sCT and corresponding patient CT (within the body contour) was 0.6 ± 14.7 (mean ± 1 SD), with a mean absolute error of 40.5 ± 8.2 Hounsfield units. Automatic contouring results were very close to the expert interobserver level (Dice similarity coefficient): prostate 0.80 ± 0.08, bladder 0.86 ± 0.12, rectum 0.84 ± 0.06, bones 0.91 ± 0.03, and body 1.00 ± 0.003. The change in monitor units between the sCT-based plans relative to the gold standard CT plan for the same dose prescription was found to be 0.3% ± 0.8%. The 3-dimensional γ pass rate was 1.00 ± 0.00 (2 mm/2%).
The MR-Sim setup and automatic sCT generation methods using standard MR sequences generates realistic contours and electron densities for prostate cancer radiation therapy dose planning and digitally reconstructed radiograph generation.
Finite element methods (FEM) are popular approaches for simulation of soft tissues with elastic or viscoelastic behavior. However, their usage in real-time applications, such as in virtual reality ...surgical training, is limited by computational cost. In this application scenario, which typically involves transportable simulators, the computing hardware severely constrains the size or the level of details of the simulated scene. To address this limitation, data-driven approaches have been suggested to simulate mechanical deformations by learning the mapping rules from FEM generated datasets. Prior data-driven approaches have ignored the physical laws of the underlying engineering problem and have consequently been restricted to simulation cases of simple hyperelastic materials where the temporal variations were effectively ignored. However, most surgical training scenarios require more complex hyperelastic models to deal with the viscoelastic properties of tissues. This type of material exhibits both viscous and elastic behaviors when subjected to external force, requiring the implementation of time-dependant state variables. Herein, we propose a deep learning method for predicting displacement fields of soft tissues with viscoelastic properties. The main contribution of this work is the use of a physics-guided loss function for the optimization of the deep learning model parameters. The proposed deep learning model is based on convolutional (CNN) and recurrent layers (LSTM) to predict spatiotemporal variations. It is augmented with a mass conservation law in the lost function to prevent the generation of physically inconsistent results. The deep learning model is trained on a set of FEM datasets that are generated from a commercially available state-of-the-art numerical neurosurgery simulator. The use of the physics-guided loss function in a deep learning model has led to a better generalization in the prediction of deformations in unseen simulation cases. Moreover, the proposed method achieves a better accuracy over the conventional CNN models, where improvements were observed in unseen tissue from 8% to 30% depending on the magnitude of external forces. It is hoped that the present investigation will help in filling the gap in applying deep learning in virtual reality simulators, hence improving their computational performance (compared to FEM simulations) and ultimately their usefulness.
•Deep learning in a commercial virtual reality simulator.•Speeding up FEM simulations using deep learning.•Deep learning for prediction of displacement fields of soft tissues.•Integration of mass-conservation law into a deep learning model.
A 2D/3D nonrigid registration method is proposed that brings a 3D centerline model of the coronary arteries into correspondence with bi-plane fluoroscopic angiograms. The registered model is overlaid ...on top of interventional angiograms to provide surgical assistance during image-guided chronic total occlusion procedures, thereby reducing the uncertainty inherent in 2D interventional images. The proposed methodology is divided into two parts: global structural alignment and local nonrigid registration. In both cases, vessel centerlines are automatically extracted from the 2D fluoroscopic images, and serve as the basis for the alignment and registration algorithms. In the first part, an energy minimization method is used to estimate a global affine transformation that aligns the centerline with the angiograms. The performance of nine general purpose optimizers has been assessed for this problem, and detailed results are presented. In the second part, a fully nonrigid registration method is proposed and used to compensate for any local shape discrepancy. This method is based on a variational framework, and uses a simultaneous matching and reconstruction process to compute a nonrigid registration. With a typical run time of less than 3 s, the algorithms are fast enough for interactive applications. Experiments on five different subjects are presented and show promising results.
Purpose:
The feasibility of radiation therapy treatment planning using substitute computed tomography (sCT) generated from magnetic resonance images (MRIs) has been demonstrated by a number of ...research groups. One challenge with an MRI-alone workflow is the accurate identification of intraprostatic gold fiducial markers, which are frequently used for prostate localization prior to each dose delivery fraction. This paper investigates a template-matching approach for the detection of these seeds in MRI.
Methods:
Two different gradient echo T1 and T2* weighted MRI sequences were acquired from fifteen prostate cancer patients and evaluated for seed detection. For training, seed templates from manual contours were selected in a spectral clustering manifold learning framework. This aids in clustering “similar” gold fiducial markers together. The marker with the minimum distance to a cluster centroid was selected as the representative template of that cluster during training. During testing, Gaussian mixture modeling followed by a Markovian model was used in automatic detection of the probable candidates. The probable candidates were rigidly registered to the templates identified from spectral clustering, and a similarity metric is computed for ranking and detection.
Results:
A fiducial detection accuracy of 95% was obtained compared to manual observations. Expert radiation therapist observers were able to correctly identify all three implanted seeds on 11 of the 15 scans (the proposed method correctly identified all seeds on 10 of the 15).
Conclusions:
An novel automatic framework for gold fiducial marker detection in MRI is proposed and evaluated with detection accuracies comparable to manual detection. When radiation therapists are unable to determine the seed location in MRI, they refer back to the planning CT (only available in the existing clinical framework); similarly, an automatic quality control is built into the automatic software to ensure that all gold seeds are either correctly detected or a warning is raised for further manual intervention.
Purpose
To demonstrate selection of a small representative subset of images from a pool of images comprising a potential atlas (PA) pelvic CT set to be used for autosegmentation of a separate target ...image set. The aim is to balance the need for the atlas set to represent anatomical diversity with the need to minimize resources required to create a high quality atlas set (such as multiobserver delineation), while retaining access to additional information available for the PA image set.
Methods
Preprocessing was performed for image standardization, followed by image registration. Clustering was used to select the subset that provided the best coverage of a target dataset as measured by postregistration image intensity similarities. Tests for clustering robustness were performed including repeated clustering runs using different starting seeds and clustering repeatedly using 90% of the target dataset chosen randomly. Comparisons of coverage of a target set (comprising 711 pelvic CT images) were made for atlas sets of five images (chosen from a PA set of 39 pelvic CT and MR images) (a) at random (averaged over 50 random atlas selections), (b) based solely on image similarities within the PA set (representing prospective atlas development), (c) based on similarities within the PA set and between the PA and target dataset (representing retrospective atlas development). Comparisons were also made to coverage provided by the entire PA set of 39 images.
Results
Exemplar selection was highly robust with exemplar selection results being unaffected by choice of starting seed with very occasional change to one of the exemplar choices when the target set was reduced. Coverage of the target set, as measured by best normalized cross‐correlation similarity of target images to any exemplar image, provided by five well‐selected atlas images (mean = 0.6497) was more similar to coverage provided by the entire PA set (mean = 0.6658) than randomly chosen atlas subsets (mean = 0.5977). This was true both of the mean values and the shape of the distributions. Retrospective selection of atlases (mean = 0.6497) provided a very small improvement over prospective atlas selection (mean = 0.6431). All differences were significant (P < 1.0E‐10).
Conclusions
Selection of a small representative image set from one dataset can be utilized to develop an atlas set for either retrospective or prospective autosegmentation of a different target dataset. The coverage provided by such a judiciously selected subset has the potential to facilitate propagation of numerous retrospectively defined structures, utilizing additional information available with multimodal imaging in the atlas set, without the need to create large atlas image sets.