Prostate cancer (PCa) is the most common cancer in men in the United States. Multiparametic magnetic resonance imaging (mp-MRI) has been explored by many researchers to targeted prostate biopsies and radiation therapy. However, assessment on mp-MRI can be subjective, development of computer-aided diagnosis systems to automatically delineate the prostate gland and the intraprostratic lesions (ILs) becomes important to facilitate with radiologists in clinical practice. In this paper, we first study the implementation of the Mask-RCNN model to segment the prostate and ILs. We trained and evaluated models on 120 patients from two different cohorts of patients. We also used 2D U-Net and 3D U-Net as benchmarks to segment the prostate and compared the model's performance. The contour variability of ILs using the algorithm was also benchmarked against the interobserver variability between two different radiation oncologists on 19 patients. Our results indicate that the Mask-RCNN model is able to reach state-of-art performance in the prostate segmentation and outperforms several competitive baselines in ILs segmentation.Automated segmentation of the prostate and screening of prostate cancer from MR images is critical for computer-aided clinical diagnosis, treatment planning and prognosis. However, the development of automatic algorithms remains challenging in several reasons. First of all, there are large variations in image quality caused by several factors at the time of image acquisition (e.g. patient motion, signal-to-noise ratio, use of an endorectal coil, Gadolinium enhancement, etc.). Second, the normal anatomy of the prostate is highly variable across patients and at different time points; and the boundaries between the prostate and surrounding structures (e.g. neurovascular bundles, bladder, rectum, seminal vessels and other soft tissues) are not always immediately clear. The prostate also shows a large variation in size and shape among different patients due to individual differences and physiological changes. Third, the presence of benign conditions such as benign prostatic hyperplasia and prostatitis may mimic the radiographic presentation of a malignancy. Contrast and pixel value of MRI also highlight a large variability in both tissue and texture information.Over the past few years, progress in image segmentation tasks has been exclusively driven by convolutional neural network (CNN) based models. Many segmentation models fall into two classes. The first class does not rely on the region proposal algorithm. Typical models in this class usually apply an encoder-decoder framework (Liou et al., 2014). The encoder network extracts representations of the image, and the decoder network reconstructs segmentation mask from the learned image representations produced by the encoder network. U-Net (Ronneberger et al., 2015), for instance, is a classic algorithm widely used in biomed-ical image segmentation tasks. Another class of models have their underlying fundamentals on region proposals such as the Mask-RCNN model,...

BackgroundStereotactic body radiation therapy (SBRT) is a proven and effective modality for treatment of hepatic primary and metastatic tumors. However, these lesions are challenging for planning and treatment execution due to natural anatomic changes associated with respiration. Magnetic resonance imaging (MRI) offers superior soft tissue contrast resolution and the ability for real-time image-guided treatment delivery and lesion tracking.ObjectiveTo evaluate the plan quality, treatment delivery, and tumor response of a set of liver SBRT cancer treatments delivered with magnetic resonance (MR)-guided radiotherapy on a MR-linear accelerator (MR-linac).MethodsTreatment data from 29 consecutive patients treated with SBRT were reviewed. All treatments were performed using a step and shoot technique to one or more liver lesions on an MR-linac platform. Patients received 45 to 50 Gy prescribed to at least 95% of the planning target volume (PTV) in five fractions except for two patients who received 27-30 Gy in three fractions. Computed tomography and MRI simulation were performed in the supine position prior to treatment in the free-breathing, end exhalation, and end inhalation breath-hold positions to determine patient tolerability and potential dosimetric advantages of each technique. Immobilization consisted of using anterior and posterior torso MRI receive coils embedded in a medium-sized vacuum cushion. Gating was performed using sagittal cine images acquired at 4 frames/second. Gating boundaries were defined in the three major axes to be 0.3 to 0.5 cm. An overlapping region of interest, defined as the percentage volume allowed outside the boundary for beam-on to occur, was set between 1 and 10%. The contoured target was assigned a 5-mm PTV expansion. Organs at risk constraints adopted by the American Association of Physicists in Medicine Task Group 101 were used during optimization.ResultsTwenty-nine patients, with a total of 34 lesions, successfully completed the prescribed treatment with minimal treatment breaks or delays. Twenty-one patients were treated at end-exhale, and six were treated at end-inhale. Two patients were treated using a free-breathing technique due to poor compliance with breath-hold instructions. The reported mean liver dose was 5.56 Gy (1.39 - 10.43; STD 2.85) and the reported mean liver volume receiving the prescribed threshold dose was 103.1 cm3 (2.9 - 236.6; STD 75.2). Follow-up imaging at one to 12 months post treatment confirmed either stable or decreased size of treated lesions in all but one patient. Toxicities were mild and included nausea/vomiting, abdominal pain and one case of bloody diarrhea. Four patients died due to complications from liver cirrhosis unrelated to radiation effect.ConclusionSBRT treatment using a gated technique on an MR-linac has been successfully demonstrated. Potential benefits of this modality include decreased liver dose leading to decreased toxicities. Further studies to identify the benefits and risks associated with MR-guided SBRT are necessary.

In women with early-stage uterine carcinosarcoma, our data suggest superior survival end points with combined RT and chemotherapy. The frequency of vaginal recurrence suggests a role for incorporating vaginal brachytherapy in the adjuvant management of this disease.

Objective: To compare survival endpoints between African American (AA) and non-AA (NAA) women with endometrial carcinoma (EC) stage I–II using a robust matching analysis. Methods and Materials: Patients were matched by stage, grade, adjuvant management (surveillance, vaginal brachytherapy or pelvic radiation treatment), age, and year of hysterectomy. Recurrence-free survival (RFS), disease-specific survival (DSS) and overall survival (OS) were calculated. Results: A total of 758 patients were included. Body mass index and Age-Adjusted Charlson comorbidity index was significantly higher in AA compared to NAA women. There were no significant differences between the AA and NAA groups in regard to 5-year RFS (94 vs. 93%), DSS (96 vs. 98%) or 5-year OS (90 vs. 92%). On multivariate analysis of survival endpoints for the entire study cohort, it was found that race (AA vs. NAA) was not a significant predictor of RFS, DSS, or OS. Grade 3 tumors and the presence of lymphovascular space invasion (LVSI) were the only 2 independent predictors of RFS and DSS, while age-adjusted Charlson comorbidity score, grade 3, stage II and the presence of LVSI were independent predictors of shorter OS. Conclusions: When matched based on the tumor stage, grade, adjuvant treatment, age, and year of surgery, our study suggests that there is no statistically significant difference in any survival endpoints between AA and NAA women with early-stage EC. Based on these data, disparities in outcome likely do not stem from uterine cancer-related causes. The increased comorbidity burden in AA women is likely a factor contributing to the racial disparity in endometrial cancer.

Introduction: Multiparametric MR imaging (mpMRI) has shown promising results in the diagnosis and localization of prostate cancer. Furthermore, mpMRI may play an important role in identifying the dominant intraprostatic lesion (DIL) for radiotherapy boost. We sought to investigate the level of correlation between dominant tumor foci contoured on various mpMRI sequences. Methods: mpMRI data from 90 patients with MR-guided biopsy-proven prostate cancer were obtained from the SPIE-AAPM-NCI Prostate MR Classification Challenge. Each case consisted of T2-weighted (T2W), apparent diffusion coefficient (ADC), and K trans images computed from dynamic contrast-enhanced sequences. All image sets were rigidly co-registered, and the dominant tumor foci were identified and contoured for each MRI sequence. Hausdorff distance (HD), mean distance to agreement (MDA), and Dice and Jaccard coefficients were calculated between the contours for each pair of MRI sequences (i.e., T2 vs. ADC, T2 vs. K trans , and ADC vs. K trans ). The voxel wise spearman correlation was also obtained between these image pairs. Results: The DILs were located in the anterior fibromuscular stroma, central zone, peripheral zone, and transition zone in 35.2, 5.6, 32.4, and 25.4% of patients, respectively. Gleason grade groups 1–5 represented 29.6, 40.8, 15.5, and 14.1% of the study population, respectively (with group grades 4 and 5 analyzed together). The mean contour volumes for the T2W images, and the ADC and K trans maps were 2.14 ± 2.1, 2.22 ± 2.2, and 1.84 ± 1.5 mL, respectively. K trans values were indistinguishable between cancerous regions and the rest of prostatic regions for 19 patients. The Dice coefficient and Jaccard index were 0.74 ± 0.13, 0.60 ± 0.15 for T2W-ADC and 0.61 ± 0.16, 0.46 ± 0.16 for T2W-K trans . The voxel-based Spearman correlations were 0.20 ± 0.20 for T2W-ADC and 0.13 ± 0.25 for T2W-K trans . Conclusions: The DIL contoured on T2W images had a high level of agreement with those contoured on ADC maps, but there was little to no quantitative correlation of these results with tumor location and Gleason grade group. Technical hurdles are yet to be solved for precision radiotherapy to target the DILs based on physiological imaging. A Boolean sum volume (BSV) incorporating all available MR sequences may be reasonable in delineating the DIL boost volume.