Uncovering prostate cancer aggressiveness signal in T2-weighted MRI through a three-reference tissues normalization technique.

Quantitative T2-weighted MRI (T2W) interpretation is impeded by the variability of acquisition-related features, such as field strength, coil type, signal amplification, and pulse sequence parameters. The main purpose of this work is to develop an automated method for prostate T2W intensity normalization. The procedure includes the following: (i) a deep learning-based network utilizing MASK R-CNN for automatic segmentation of three reference tissues: gluteus maximus muscle, femur, and bladder; (ii) fitting a spline function between average intensities in these structures and reference values; and (iii) using the function to transform all T2W intensities. The T2W distributions in the prostate cancer regions of interest (ROIs) and normal appearing prostate tissue (NAT) were compared before and after normalization using Student's t-test. The ROIs' T2W associations with the Gleason Score (GS), Decipher genomic score, and a three-tier prostate cancer risk were evaluated with Spearman's correlation coefficient (rS ). T2W differences in indolent and aggressive prostate cancer lesions were also assessed. The MASK R-CNN was trained with manual contours from 32 patients. The normalization procedure was applied to an independent MRI dataset from 83 patients. T2W differences between ROIs and NAT significantly increased after normalization. T2W intensities in 231 biopsy ROIs were significantly negatively correlated with GS (rS = -0.21, p = 0.001), Decipher (rS = -0.193, p = 0.003), and three-tier risk (rS = -0.235, p < 0.001). The average T2W intensities in the aggressive ROIs were significantly lower than in the indolent ROIs after normalization. In conclusion, the automated triple-reference tissue normalization method significantly improved the discrimination between prostate cancer and normal prostate tissue. In addition, the normalized T2W intensities of cancer exhibited a significant association with tumor aggressiveness. By improving the quantitative utilization of the T2W in the assessment of prostate cancer on MRI, the new normalization method represents an important advance over clinical protocols that do not include sequences for the measurement of T2 relaxation times.

Clinical-Genomic Risk Group Classification of Suspicious Lesions on Prostate Multiparametric-MRI.

The utilization of multi-parametric MRI (mpMRI) in clinical decisions regarding prostate cancer patients' management has recently increased. After biopsy, clinicians can assess risk using National Comprehensive Cancer Network (NCCN) risk stratification schema and commercially available genomic classifiers, such as Decipher. We built radiomics-based models to predict lesions/patients at low risk prior to biopsy based on an established three-tier clinical-genomic classification system. Radiomic features were extracted from regions of positive biopsies and Normally Appearing Tissues (NAT) on T2-weighted and Diffusion-weighted Imaging. Using only clinical information available prior to biopsy, five models for predicting low-risk lesions/patients were evaluated, based on: 1: Clinical variables; 2: Lesion-based radiomic features; 3: Lesion and NAT radiomics; 4: Clinical and lesion-based radiomics; and 5: Clinical, lesion and NAT radiomic features. Eighty-three mpMRI exams from 78 men were analyzed. Models 1 and 2 performed similarly (Area under the receiver operating characteristic curve were 0.835 and 0.838, respectively), but radiomics significantly improved the lesion-based performance of the model in a subset analysis of patients with a negative Digital Rectal Exam (DRE). Adding normal tissue radiomics significantly improved the performance in all cases. Similar patterns were observed on patient-level models. To the best of our knowledge, this is the first study to demonstrate that machine learning radiomics-based models can predict patients' risk using combined clinical-genomic classification.

A Deep Learning Approach for Automatic Segmentation during Daily MRI-Linac Radiotherapy of Glioblastoma.

Glioblastoma changes during chemoradiotherapy are inferred from high-field MRI before and after treatment but are rarely investigated during radiotherapy. The purpose of this study was to develop a deep learning network to automatically segment glioblastoma tumors on daily treatment set-up scans from the first glioblastoma patients treated on MRI-linac. Glioblastoma patients were prospectively imaged daily during chemoradiotherapy on 0.35T MRI-linac. Tumor and edema (tumor lesion) and resection cavity kinetics throughout the treatment were manually segmented on these daily MRI. Utilizing a convolutional neural network, an automatic segmentation deep learning network was built. A nine-fold cross-validation schema was used to train the network using 80:10:10 for training, validation, and testing. Thirty-six glioblastoma patients were imaged pre-treatment and 30 times during radiotherapy (n = 31 volumes, total of 930 MRIs). The average tumor lesion and resection cavity volumes were 94.56 ± 64.68 cc and 72.44 ± 35.08 cc, respectively. The average Dice similarity coefficient between manual and auto-segmentation for tumor lesion and resection cavity across all patients was 0.67 and 0.84, respectively. This is the first brain lesion segmentation network developed for MRI-linac. The network performed comparably to the only other published network for auto-segmentation of post-operative glioblastoma lesions. Segmented volumes can be utilized for adaptive radiotherapy and propagated across multiple MRI contrasts to create a prognostic model for glioblastoma based on multiparametric MRI.

Generative Adversarial Networks Can Create High Quality Artificial Prostate Cancer Magnetic Resonance Images.

The recent integration of open-source data with machine learning models, especially in the medical field, has opened new doors to studying disease progression and/or regression. However, the ability to use medical data for machine learning approaches is limited by the specificity of data for a particular medical condition. In this context, the most recent technologies, like generative adversarial networks (GANs), are being looked upon as a potential way to generate high-quality synthetic data that preserve the clinical variability of a condition. However, despite some success, GAN model usage remains largely minimal when depicting the heterogeneity of a disease such as prostate cancer. Previous studies from our group members have focused on automating the quantitative multi-parametric magnetic resonance imaging (mpMRI) using habitat risk scoring (HRS) maps on the prostate cancer patients in the BLaStM trial. In the current study, we aimed to use the images from the BLaStM trial and other sources to train the GAN models, generate synthetic images, and validate their quality. In this context, we used T2-weighted prostate MRI images as training data for Single Natural Image GANs (SinGANs) to make a generative model. A deep learning semantic segmentation pipeline trained the model to segment the prostate boundary on 2D MRI slices. Synthetic images with a high-level segmentation boundary of the prostate were filtered and used in the quality control assessment by participating scientists with varying degrees of experience (more than ten years, one year, or no experience) to work with MRI images. Results showed that the most experienced participating group correctly identified conventional vs. synthetic images with 67% accuracy, the group with one year of experience correctly identified the images with 58% accuracy, and the group with no prior experience reached 50% accuracy. Nearly half (47%) of the synthetic images were mistakenly evaluated as conventional. Interestingly, in a blinded quality assessment, a board-certified radiologist did not significantly differentiate between conventional and synthetic images in the context of the mean quality of synthetic and conventional images. Furthermore, to validate the usability of the generated synthetic images from prostate cancer MRIs, we subjected these to anomaly detection along with the original images. Importantly, the success rate of anomaly detection for quality control-approved synthetic data in phase one corresponded to that of the conventional images. In sum, this study shows promise that high-quality synthetic images from MRIs can be generated using GANs. Such an AI model may contribute significantly to various clinical applications which involve supervised machine-learning approaches.

Simulated Adaptive Radiotherapy for Shrinking Glioblastoma Resection Cavities on a Hybrid MRI-Linear Accelerator.

During radiation therapy (RT) of glioblastoma, daily MRI with combination MRI-linear accelerator (MRI-Linac) systems has demonstrated significant anatomic changes, including evolving post-surgical cavity shrinkage. Cognitive function RT for brain tumors is correlated with radiation doses to healthy brain structures, especially the hippocampi. Therefore, this study investigates whether adaptive planning to the shrinking target could reduce normal brain RT dose with the goal of improving post-RT function. We evaluated 10 glioblastoma patients previously treated on a 0.35T MRI-Linac with a prescription of 60 Gy delivered in 30 fractions over six weeks without adaptation ("static plan") with concurrent temozolomide chemotherapy. Six weekly plans were created per patient. Reductions in the radiation dose to uninvolved hippocampi (maximum and mean) and brain (mean) were observed for weekly adaptive plans. The dose (Gy) to the hippocampi for static vs. weekly adaptive plans were, respectively: max 21 ± 13.7 vs. 15.2 ± 8.2 (p = 0.003) and mean 12.5 ± 6.7 vs. 8.4 ± 4.0 (p = 0.036). The mean brain dose was 20.6 ± 6.0 for static planning vs. 18.7 ± 6.8 for weekly adaptive planning (p = 0.005). Weekly adaptive re-planning has the potential to spare the brain and hippocampi from high-dose radiation, possibly reducing the neurocognitive side effects of RT for eligible patients.

Longitudinal Changes and Predictive Value of Multiparametric MRI Features for Prostate Cancer Patients Treated with MRI-Guided Lattice Extreme Ablative Dose (LEAD) Boost Radiotherapy.

We investigated the longitudinal changes in multiparametric MRI (mpMRI) (T2-weighted, Apparent Diffusion Coefficient (ADC), and Dynamic Contrast Enhanced (DCE-)MRI) of prostate cancer patients receiving Lattice Extreme Ablative Dose (LEAD) radiotherapy (RT) and the capability of their imaging features to predict RT outcome based on endpoint biopsies. Ninety-five mpMRI exams from 25 patients, acquired pre-RT and at 3-, 9-, and 24-months post-RT were analyzed. MRI/Ultrasound-fused biopsies were acquired pre- and at two-years post-RT (endpoint). Five regions of interest (ROIs) were analyzed: Gross tumor volume (GTV), normally-appearing tissue (NAT) and peritumoral volume in both peripheral (PZ) and transition (TZ) zones. Diffusion and perfusion radiomics features were extracted from mpMRI and compared before and after RT using two-tailed Student t-tests. Selected features at the four scan points and their differences (Δ radiomics) were used in multivariate logistic regression models to predict the endpoint biopsy positivity. Baseline ADC values were significantly different between GTV, NAT-PZ, and NAT-TZ (p-values < 0.005). Pharmaco-kinetic features changed significantly in the GTV at 3-month post-RT compared to baseline. Several radiomics features at baseline and three-months post-RT were significantly associated with endpoint biopsy positivity and were used to build models with high predictive power of this endpoint (AUC = 0.98 and 0.89, respectively). Our study characterized the RT-induced changes in perfusion and diffusion. Quantitative imaging features from mpMRI show promise as being predictive of endpoint biopsy positivity.

Deep Learning for Per-Fraction Automatic Segmentation of Gross Tumor Volume (GTV) and Organs at Risk (OARs) in Adaptive Radiotherapy of Cervical Cancer.

Background/Hypothesis: MRI-guided online adaptive radiotherapy (MRI-g-OART) improves target coverage and organs-at-risk (OARs) sparing in radiation therapy (RT). For patients with locally advanced cervical cancer (LACC) undergoing RT, changes in bladder and rectal filling contribute to large inter-fraction target volume motion. We hypothesized that deep learning (DL) convolutional neural networks (CNN) can be trained to accurately segment gross tumor volume (GTV) and OARs both in planning and daily fractions' MRI scans. Materials/Methods: We utilized planning and daily treatment fraction setup (RT-Fr) MRIs from LACC patients, treated with stereotactic body RT to a dose of 45-54 Gy in 25 fractions. Nine structures were manually contoured. MASK R-CNN network was trained and tested under three scenarios: (i) Leave-one-out (LOO), using the planning images of N- 1 patients for training; (ii) the same network, tested on the RT-Fr MRIs of the "left-out" patient, (iii) including the planning MRI of the "left-out" patient as an additional training sample, and tested on RT-Fr MRIs. The network performance was evaluated using the Dice Similarity Coefficient (DSC) and Hausdorff distances. The association between the structures' volume and corresponding DSCs was investigated using Pearson's Correlation Coefficient, r. Results: MRIs from fifteen LACC patients were analyzed. In the LOO scenario the DSC for Rectum, Femur, and Bladder was >0.8, followed by the GTV, Uterus, Mesorectum and Parametrium (0.6-0.7). The results for Vagina and Sigmoid were suboptimal. The performance of the network was similar for most organs when tested on RT-Fr MRI. Including the planning MRI in the training did not improve the segmentation of the RT-Fr MRI. There was a significant correlation between the average organ volume and the corresponding DSC (r = 0.759, p = 0.018). Conclusion: We have established a robust workflow for training MASK R-CNN to automatically segment GTV and OARs in MRI-g-OART of LACC. Albeit the small number of patients in this pilot project, the network was trained to successfully identify several structures while challenges remain, especially in relatively small organs. With the increase of the LACC cases, the performance of the network will improve. A robust auto-contouring tool would improve workflow efficiency and patient tolerance of the OART process.

Segmentation of prostate and prostate zones using deep learning : A multi-MRI vendor analysis.

PURPOSE: Develop a deep-learning-based segmentation algorithm for prostate and its peripheral zone (PZ) that is reliable across multiple MRI vendors. METHODS: This is a retrospective study. The dataset consisted of 550 MRIs (Siemens-330, General Electric[GE]-220). A multistream 3D convolutional neural network is used for automatic segmentation of the prostate and its PZ using T2-weighted (T2-w) MRI. Prostate and PZ were manually contoured on axial T2­w. The network uses axial, coronal, and sagittal T2­w series as input. The preprocessing of the input data includes bias correction, resampling, and image normalization. A dataset from two MRI vendors (Siemens and GE) is used to test the proposed network. Six different models were trained, three for the prostate and three for the PZ. Of the three, two were trained on data from each vendor separately, and a third (Combined) on the aggregate of the datasets. The Dice coefficient (DSC) is used to compare the manual and predicted segmentation. RESULTS: For prostate segmentation, the Combined model obtained DSCs of 0.893 ± 0.036 and 0.825 ± 0.112 (mean ± standard deviation) on Siemens and GE, respectively. For PZ, the best DSCs were from the Combined model: 0.811 ± 0.079 and 0.788 ± 0.093. While the Siemens model underperformed on the GE dataset and vice versa, the Combined model achieved robust performance on both datasets. CONCLUSION: The proposed network has a performance comparable to the interexpert variability for segmenting the prostate and its PZ. Combining images from different MRI vendors on the training of the network is of paramount importance for building a universal model for prostate and PZ segmentation.

Automatic Detection of Prostate Tumor Habitats using Diffusion MRI.

A procedure for identification of optimal Apparent Diffusion Coefficient (ADC) thresholds for automatic delineation of prostatic lesions with restricted diffusion at differing risk for cancer was developed. The relationship between the size of the identified Volumes of Interest (VOIs) and Gleason Score (GS) was evaluated. Patients with multiparametric (mp)MRI, acquired prior to radical prostatectomy (RP) (n = 18), mpMRI-ultrasound fused (MRI-US) (n = 21) or template biopsies (n = 139) were analyzed. A search algorithm, spanning ADC thresholds in 50 µm2/s increments, determined VOIs that were matched to RP tumor nodules. Three ADC thresholds for both peripheral zone (PZ) and transition zone (TZ) were identified for estimation of VOIs at low, intermediate, and high risk of prostate cancer. The determined ADC thresholds for low, intermediate and high risk in PZ/TZ were: 900/800; 1100/850; and 1300/1050 µm2/s. The correlation coefficients between the size of the high/intermediate/low risk VOIs and GS in the three cohorts were 0.771/0.778/0.369, 0.561/0.457/0.355 and 0.423/0.441/0.36 (p < 0.05). Low risk VOIs mapped all RP lesions; area under the curve (AUC) for intermediate risk VOIs to discriminate GS6 vs GS ≥ 7 was 0.852; for high risk VOIs to discriminate GS6,7 vs GS ≥ 8 was 0.952. In conclusion, the automatically delineated volumes in the prostate with restricted diffusion were found to strongly correlate with cancer aggressiveness.

Magnetic resonance imaging (MRI)-based radiomics for prostate cancer radiotherapy.

In radiotherapy (RT) of prostate cancer, dose escalation has been shown to reduce biochemical failure. Dose escalation only to determinate prostate tumor habitats has the potential to improve tumor control with less toxicity than when the entire prostate is dose escalated. Other issues in the treatment of the RT patient include the choice of the RT technique (hypo- or standard fractionation) and the use and length of concurrent/adjuvant androgen deprivation therapy (ADT). Up to 50% of high-risk men demonstrate biochemical failure suggesting that additional strategies for defining and treating patients based on improved risk stratification are required. The use of multiparametric MRI (mpMRI) is rapidly gaining momentum in the management of prostate cancer because of its improved diagnostic potential and its ability to combine functional and anatomical information. Currently, the Prostate Imaging, Reporting and Diagnosis System (PIRADS) is the standard of care for region of interest (ROI) identification and risk classification. However, PIRADS was not designed for 3D tumor volume delineation; there is a large degree of subjectivity and PIRADS does not accurately and reproducibly elucidate inter- and intra-lesional spatial heterogeneity. "Radiomics", as it refers to the extraction and analysis of large number of advanced quantitative radiological features from medical images using high throughput methods, is perfectly suited as an engine to effectively sift through the multiple series of prostate mpMRI sequences and quantify regions of interest. The radiomic efforts can be summarized in two main areas: (I) detection/segmentation of the suspicious lesion; and (II) assessment of the aggressiveness of prostate cancer. As related to RT, the goal of the latter is in particular to identify patients at high risk for metastatic disease; and the aim of the former is to identify and segment cancerous lesions and thus provide targets for radiation boost. The article is structured as follows: first, we describe the radiomic approach; and second, we discuss the radiomic pipeline as tailored for RT of prostate cancer. In this process we summarize the current efforts and progress in integrating mpMRI radiomics into the radiotherapeutic management of prostate cancer with emphasis placed on its role in treatment target definition, treatment plan strategizing, and prognostic assessment. The described concepts, methods and tools are not currently applicable to the radiation oncology practice outside of the research setting. More data are required in the form of clinical trials to assess the robustness of radiomics-based predictive models, and to maximize the efficacy of these models.

An Automated Multiparametric MRI Quantitative Imaging Prostate Habitat Risk Scoring System for Defining External Beam Radiation Therapy Boost Volumes.

PURPOSE: To develop a prostate tumor habitat risk scoring (HRS) system based on multiparametric magnetic resonance imaging (mpMRI) referenced to prostatectomy Gleason score (GS) for automatic delineation of gross tumor volumes. A workflow for integration of HRS into radiation therapy boost volume dose escalation was developed in the framework of a phase 2 randomized clinical trial (BLaStM). METHODS AND MATERIALS: An automated quantitative mpMRI-based 10-point pixel-by-pixel method was optimized to prostatectomy GSs and volumes using referenced dynamic contrast-enhanced and apparent diffusion coefficient sequences. The HRS contours were migrated to the planning computed tomography scan for boost volume generation. RESULTS: There were 51 regions of interest in 12 patients who underwent radical prostatectomy (26 with GS ≥7 and 25 with GS 6). The resultant heat maps showed inter- and intratumoral heterogeneity. The HRS6 level was significantly associated with radical prostatectomy regions of interest (slope 1.09, r = 0.767; P < .0001). For predicting the likelihood of cancer, GS ≥7 and GS ≥8 HRS6 area under the curve was 0.718, 0.802, and 0.897, respectively. HRS was superior to the Prostate Imaging, Reporting and Diagnosis System 4/5 classification, wherein the area under the curve was 0.62, 0.64, and 0.617, respectively (difference with HR6, P < .0001). HRS maps were created for the first 37 assessable patients on the BLaStM trial. There were an average of 1.38 habitat boost volumes per patient at a total boost volume average of 3.6 cm3. CONCLUSIONS: An automated quantitative mpMRI-based method was developed to objectively guide dose escalation to high-risk habitat volumes based on prostatectomy GS.

Analysis of Magnetic Resonance Image Signal Fluctuations Acquired During MR-Guided Radiotherapy.

Magnetic resonance-guided radiotherapy (MRgRT) is a new and evolving treatment modality that allows unprecedented visualization of the tumor and surrounding anatomy. MRgRT includes daily 3D magnetic resonance imaging (MRI) for setup and rapidly repeated near real-time MRI scans during treatment for target tracking. One of the more exciting potential benefits of MRgRT is the ability to analyze serial MRIs to monitor treatment response or predict outcomes. A typical radiation treatment (RT) over the span of 10-15 minutes on the MRIdian system (ViewRay, Cleveland, OH) yields thousands of "cine" images, each acquired in 250 ms. This unique data allows for a glimpse in image intensity changes during RT delivery. In this report, we analyze cine images from a single fraction RT of a glioblastoma patient on the ViewRay platform in order to characterize the dynamic signal changes occurring during RT therapy. The individual frames in the cines were saved into DICOM format and read into an MIM image analysis platform (MIM Software, Cleveland, OH) as a time series. The three possible states of the three Cobalt-60 radiation sources-OFF, READY, and ON-were also recorded. An in-house Java plugin for MIM was created in order to perform principal component analysis (PCA) on each of the datasets. The analysis resulted in first PC, related to monotonous signal increase over the course of the treatment fraction. We found several distortion patterns in the data that we postulate result from the perturbation of the magnetic field due to the moving metal parts in the platform while treatment was being administered. The largest variations were detected when all Cobalt-60 sources were OFF. During this phase of the treatment, the gantry and multi-leaf collimators (MLCs) are moving. Conversely, when all Cobalt-60 sources were in the ON position, the image signal fluctuations were minimal, relating to very little mechanical motion. At this phase, the gantry, the MLCs, and sources are fixed in their positions. These findings were confirmed in a study with the daily quality assurance (QA) phantom. While the identified variations were not related to physiological processes, our findings confirm the sensitivity of the developed approach to identify very small fluctuations. Relating these variations to the physical changes that occur during treatment shows the methodical ability of the technique to uncover their underlying sources.

Classification of suspicious lesions on prostate multiparametric MRI using machine learning.

We present a radiomics-based approach developed for the SPIE-AAPM-NCI PROSTATEx challenge. The task was to classify clinically significant prostate cancer in multiparametric (mp) MRI. Data consisted of a "training dataset" (330 suspected lesions from 204 patients) and a "test dataset" (208 lesions/140 patients). All studies included T2-weighted (T2-W), proton density-weighted, dynamic contrast enhanced, and diffusion-weighted imaging. Analysis of the images was performed using the MIM imaging platform (MIM Software, Cleveland, Ohio). Prostate and peripheral zone contours were manually outlined on the T2-W images. A workflow for rigid fusion of the aforementioned images to T2-W was created in MIM. The suspicious lesion was outlined using the high b-value image. Intensity and texture features were extracted on four imaging modalities and characterized using nine histogram descriptors: 10%, 25%, 50%, 75%, 90%, mean, standard deviation, kurtosis, and skewness (216 features). Three classification methods were used: classification and regression trees (CART), random forests, and adaptive least absolute shrinkage and selection operator (LASSO). In the held out by the organizers test dataset, the areas under the curve (AUCs) were: 0.82 (random forests), 0.76 (CART), and 0.76 (adaptive LASSO). AUC of 0.82 was the fourth-highest score of 71 entries (32 teams) and the highest for feature-based methods.