Where Does Auto-Segmentation for Brain Metastases Radiosurgery Stand Today?

Detection and segmentation of brain metastases (BMs) play a pivotal role in diagnosis, treatment planning, and follow-up evaluations for effective BM management. Given the rising prevalence of BM cases and its predominantly multiple onsets, automated segmentation is becoming necessary in stereotactic radiosurgery. It not only alleviates the clinician's manual workload and improves clinical workflow efficiency but also ensures treatment safety, ultimately improving patient care. Recent strides in machine learning, particularly in deep learning (DL), have revolutionized medical image segmentation, achieving state-of-the-art results. This review aims to analyze auto-segmentation strategies, characterize the utilized data, and assess the performance of cutting-edge BM segmentation methodologies. Additionally, we delve into the challenges confronting BM segmentation and share insights gleaned from our algorithmic and clinical implementation experiences.

Automated contouring, treatment planning, and quality assurance for VMAT craniospinal irradiation (VMAT-CSI).

Purpose: Create a comprehensive automated solution for pediatric and adult VMAT-CSI including contouring, planning, and plan check to reduce planning time and improve plan quality. Methods: Seventy-seven previously treated CSI patients (age, 2-67 years) were used for creation of an auto-contouring model to segment 25 organs at risk (OARs). The auto-contoured OARs were evaluated using the Dice Similarity Coefficient (DSC), 95% Hausdorff Distance (HD95), and a qualitative ranking by one physician and one physicist (scale: 1-acceptable, 2-minor edits, 3-major edits). The auto-planning script was developed using the Varian Eclipse Scripting API and tested with 20 patients previously treated with either low-dose VMAT-CSI (12 Gy) or high-dose VMAT-CSI (36 Gy + 18 Gy boost). Clinically relevant metrics, planning time, and blinded physician review were used to evaluate significance of differences between the auto and manual plans. Finally, the plan preparation for treatment and plan check processes were automated to improve efficiency and safety of VMAT-CSI. Results: The auto-contours achieved an average DSC of 0.71 ± 0.15, HD95 of 4.81 ± 4.68, and reviewers' ranking of 1.22 ± 0.39, indicating close to "acceptable-as-is" contours. Compared to the manual CSI plans, the auto-plans for both dose regimens achieved statistically significant reductions in body V50% and Dmean for parotids, submandibular, and thyroid glands. The variance in the dosimetric parameters decreased for the auto-plans as compared to the manual plans indicating better plan consistency. From the blinded review, the auto-plans were marked as equivalent or superior to the manual-plans 88.3% of the time. The required time for the auto-contouring and planning was consistently between 1-2 hours compared to an estimated 5-6 hours for manual contouring and planning. Conclusions: Reductions in contouring and planning time without sacrificing plan quality were obtained using the developed auto-planning process. The auto-planning scripts and documentation will be made freely available to other institutions and clinics.

Automating the Treatment Planning Process for Volumetric Modulated Arc Therapy Craniospinal Irradiation.

PURPOSE: The purpose of this work is to develop a method to automate the treatment planning process of craniospinal irradiation (CSI) using volumetric modulated arc therapy. METHODS AND MATERIALS: Two scripts were developed using the Eclipse Scripting Application Programming Interface to perform auto-plan preparation and optimization. Ten patients (age, 5-44 years) previously treated at our institution with low dose volumetric modulated arc therapy CSI (prescription of 12 Gy) before total body irradiation were selected to evaluate the efficacy of the proposed auto-planning process. Paired t tests compared the dosimetric indices of the auto-plans to the manually generated clinical plans. All plans were normalized to 95% of planning target volume (PTV) coverage with the prescription dose. Two physicians and one physicist were asked to evaluate the manual plans and auto-plans of each patient in a blinded retrospective review and to indicate clinical acceptability and which plans were preferred for treatment. RESULTS: Compared with the manual CSI plans, the auto plans obtained significant reductions in Dmean to the parotids, submandibular glands, larynx, thyroid, and significant reduction in the plan PTV Dmax and D0.03 cc. The standard deviation range of the dosimetric parameters was greatly reduced for auto plans (range, 0.1-1.3 Gy) relative to manual plans (range, 0.4-5.9 Gy) indicating better plan consistency. Among the 10 patients, the auto-plans were preferred over the manual plans 90% of the time by the reviewing experts. The required time for auto-planning was approximately 1 hour compared with estimated 4 or more hours for manual planning. CONCLUSIONS: Reductions in planning time without sacrifices in plan quality were obtained using the auto-planning process compared with manual planning. Variation in plan quality was also reduced. The auto-planning scripts will be made freely available to other institutions and clinics.

Biology-guided deep learning predicts prognosis and cancer immunotherapy response.

Substantial progress has been made in using deep learning for cancer detection and diagnosis in medical images. Yet, there is limited success on prediction of treatment response and outcomes, which has important implications for personalized treatment strategies. A significant hurdle for clinical translation of current data-driven deep learning models is lack of interpretability, often attributable to a disconnect from the underlying pathobiology. Here, we present a biology-guided deep learning approach that enables simultaneous prediction of the tumor immune and stromal microenvironment status as well as treatment outcomes from medical images. We validate the model for predicting prognosis of gastric cancer and the benefit from adjuvant chemotherapy in a multi-center international study. Further, the model predicts response to immune checkpoint inhibitors and complements clinically approved biomarkers. Importantly, our model identifies a subset of mismatch repair-deficient tumors that are non-responsive to immunotherapy and may inform the selection of patients for combination treatments.

Non-invasive tumor microenvironment evaluation and treatment response prediction in gastric cancer using deep learning radiomics.

The tumor microenvironment (TME) plays a critical role in disease progression and is a key determinant of therapeutic response in cancer patients. Here, we propose a noninvasive approach to predict the TME status from radiological images by combining radiomics and deep learning analyses. Using multi-institution cohorts of 2,686 patients with gastric cancer, we show that the radiological model accurately predicted the TME status and is an independent prognostic factor beyond clinicopathologic variables. The model further predicts the benefit from adjuvant chemotherapy for patients with localized disease. In patients treated with checkpoint blockade immunotherapy, the model predicts clinical response and further improves predictive accuracy when combined with existing biomarkers. Our approach enables noninvasive assessment of the TME, which opens the door for longitudinal monitoring and tracking response to cancer therapy. Given the routine use of radiologic imaging in oncology, our approach can be extended to many other solid tumor types.

Adaptive Region-Specific Loss for Improved Medical Image Segmentation.

Defining the loss function is an important part of neural network design and critically determines the success of deep learning modeling. A significant shortcoming of the conventional loss functions is that they weight all regions in the input image volume equally, despite the fact that the system is known to be heterogeneous (i.e., some regions can achieve high prediction performance more easily than others). Here, we introduce a region-specific loss to lift the implicit assumption of homogeneous weighting for better learning. We divide the entire volume into multiple sub-regions, each with an individualized loss constructed for optimal local performance. Effectively, this scheme imposes higher weightings on the sub-regions that are more difficult to segment, and vice versa. Furthermore, the regional false positive and false negative errors are computed for each input image during a training step and the regional penalty is adjusted accordingly to enhance the overall accuracy of the prediction. Using different public and in-house medical image datasets, we demonstrate that the proposed regionally adaptive loss paradigm outperforms conventional methods in the multi-organ segmentations, without any modification to the neural network architecture or additional data preparation.

Stratified assessment of an FDA-cleared deep learning algorithm for automated detection and contouring of metastatic brain tumors in stereotactic radiosurgery.

PURPOSE: Artificial intelligence-based tools can be leveraged to improve detection and segmentation of brain metastases for stereotactic radiosurgery (SRS). VBrain by Vysioneer Inc. is a deep learning algorithm with recent FDA clearance to assist in brain tumor contouring. We aimed to assess the performance of this tool by various demographic and clinical characteristics among patients with brain metastases treated with SRS. MATERIALS AND METHODS: We randomly selected 100 patients with brain metastases who underwent initial SRS on the CyberKnife from 2017 to 2020 at a single institution. Cases with resection cavities were excluded from the analysis. Computed tomography (CT) and axial T1-weighted post-contrast magnetic resonance (MR) image data were extracted for each patient and uploaded to VBrain. A brain metastasis was considered "detected" when the VBrain- "predicted" contours overlapped with the corresponding physician contours ("ground-truth" contours). We evaluated performance of VBrain against ground-truth contours using the following metrics: lesion-wise Dice similarity coefficient (DSC), lesion-wise average Hausdorff distance (AVD), false positive count (FP), and lesion-wise sensitivity (%). Kruskal-Wallis tests were performed to assess the relationships between patient characteristics including sex, race, primary histology, age, and size and number of brain metastases, and performance metrics such as DSC, AVD, FP, and sensitivity. RESULTS: We analyzed 100 patients with 435 intact brain metastases treated with SRS. Our cohort consisted of patients with a median number of 2 brain metastases (range: 1 to 52), median age of 69 (range: 19 to 91), and 50% male and 50% female patients. The primary site breakdown was 56% lung, 10% melanoma, 9% breast, 8% gynecological, 5% renal, 4% gastrointestinal, 2% sarcoma, and 6% other, while the race breakdown was 60% White, 18% Asian, 3% Black/African American, 2% Native Hawaiian or other Pacific Islander, and 17% other/unknown/not reported. The median tumor size was 0.112 c.c. (range: 0.010-26.475 c.c.). We found mean lesion-wise DSC to be 0.723, mean lesion-wise AVD to be 7.34% of lesion size (0.704 mm), mean FP count to be 0.72 tumors per case, and lesion-wise sensitivity to be 89.30% for all lesions. Moreover, mean sensitivity was found to be 99.07%, 97.59%, and 96.23% for lesions with diameter equal to and greater than 10 mm, 7.5 mm, and 5 mm, respectively. No other significant differences in performance metrics were observed across demographic or clinical characteristic groups. CONCLUSION: In this study, a commercial deep learning algorithm showed promising results in segmenting brain metastases, with 96.23% sensitivity for metastases with diameters of 5 mm or higher. As the software is an assistive AI, future work of VBrain integration into the clinical workflow can provide further clinical and research insights.

Learning image representations for content-based image retrieval of radiotherapy treatment plans.

Objective.In this work, we propose a content-based image retrieval (CBIR) method for retrieving dose distributions of previously planned patients based on anatomical similarity. Retrieved dose distributions from this method can be incorporated into automated treatment planning workflows in order to streamline the iterative planning process. As CBIR has not yet been applied to treatment planning, our work seeks to understand which current machine learning models are most viable in this context.Approach.Our proposed CBIR method trains a representation model that produces latent space embeddings of a patient's anatomical information. The latent space embeddings of new patients are then compared against those of previous patients in a database for image retrieval of dose distributions. All source code for this project is available on github.Main results.The retrieval performance of various CBIR methods is evaluated on a dataset consisting of both publicly available image sets and clinical image sets from our institution. This study compares various encoding methods, ranging from simple autoencoders to more recent Siamese networks like SimSiam, and the best performance was observed for the multitask Siamese network.Significance.Our current results demonstrate that excellent image retrieval performance can be obtained through slight changes to previously developed Siamese networks. We hope to integrate CBIR into automated planning workflow in future works.

Leveraging data-driven self-consistency for high-fidelity gene expression recovery.

Single cell RNA sequencing is a promising technique to determine the states of individual cells and classify novel cell subtypes. In current sequence data analysis, however, genes with low expressions are omitted, which leads to inaccurate gene counts and hinders downstream analysis. Recovering these omitted expression values presents a challenge because of the large size of the data. Here, we introduce a data-driven gene expression recovery framework, referred to as self-consistent expression recovery machine (SERM), to impute the missing expressions. Using a neural network, the technique first learns the underlying data distribution from a subset of the noisy data. It then recovers the overall expression data by imposing a self-consistency on the expression matrix, thus ensuring that the expression levels are similarly distributed in different parts of the matrix. We show that SERM improves the accuracy of gene imputation with orders of magnitude enhancement in computational efficiency in comparison to the state-of-the-art imputation techniques.

Local control of brain metastases with osimertinib alone in patients with EGFR-mutant non-small cell lung cancer.

PURPOSE: Although osimertinib has excellent intracranial activity in metastatic non-small cell lung cancer (NSCLC) with exon 19 deletion or L858R EGFR alterations, measures of local control of brain metastases are less well-reported. We describe lesion-level outcomes of brain metastases treated with osimertinib alone. METHODS: We retrospectively reviewed patients with EGFR-mutant NSCLC with untreated brain metastasis measuring ≥ 5 mm at the time of initiating osimertinib. Cumulative incidence of local recurrence in brain (LRiB) was calculated with death as a competing risk, and univariable and multivariable analyses were conducted to identify factors associated with LRiB. RESULTS: We included 284 brain metastases from 37 patients. Median follow-up was 20.1 months. On initial MRI after starting osimertinib, patient-level response was complete response (CR) in 11 (15%), partial response (PR) in 33 (45%), stable disease (SD) in 18 (25%) and progressive disease (PD) in 11 (15%). The 1-year cumulative incidence of LRiB was 14% (95% CI 9.9-17.9) and was significantly different in patients with a CR (0%), PR (4%), and SD (11%; p = 0.02). Uncontrolled primary tumor (adjusted hazard ratio [aHR] 3.78, 95% CI 1.87-7.66; p < 0.001), increasing number of prior systemic therapies (aHR 2.12, 95% CI 1.49-3.04; p < 0.001), and higher ECOG score (aHR 7.8, 95% CI 1.99-31.81; p = 0.003) were associated with LRiB. CONCLUSIONS: Although 1-year cumulative incidence of LRiB is < 4% with a CR or PR, 1-year cumulative incidence of LRiB is over 10% for patients with less than a PR to osimertinib on initial MRI. These patients should be followed closely for need for additional treatment such as stereotactic radiosurgery.

Genotype/Phenotype Interactions and First Steps Toward Targeted Therapy for Sphingosine Phosphate Lyase Insufficiency Syndrome.

Sphingosine-1-phosphate lyase insufficiency syndrome (SPLIS) is a rare metabolic disorder caused by a deficiency in sphingosine-1-phosphate lyase (SPL), the final enzyme in the sphingolipid degradative pathway. Inactivating mutations of SGPL1-the gene encoding SPL-lead to a deficiency of its downstream products, and buildup of sphingolipid intermediates, including its bioactive substrate, sphingosine-1-phosphate (S1P), the latter causing lymphopenia, a hallmark of the disease. Other manifestations of SPLIS include nephrotic syndrome, neuronal defects, and adrenal insufficiency, but their pathogenesis remains unknown. In this report, we describe the correlation between SGPL1 genotypes, age at diagnosis, and patient outcome. Vitamin B6 serves as a cofactor for SPL. B6 supplementation may aid some SPLIS patients by overcoming poor binding kinetics and promoting proper folding and stability of mutant SPL proteins. However, this approach remains limited to patients with a susceptible allele. Gene therapy represents a potential targeted therapy for SPLIS patients harboring B6-unresponsive missense mutations, truncations, deletions, and splice-site mutations. When Sgpl1 knockout (SPLKO) mice that model SPLIS were treated with adeno-associated virus (AAV)-mediated SGPL1 gene therapy, they showed profound improvement in survival and kidney and neurological function compared to untreated SPLKO mice. Thus, gene therapy appears promising as a universal, potentially curative treatment for SPLIS.

Efficacy of AAV9-mediated SGPL1 gene transfer in a mouse model of S1P lyase insufficiency syndrome.

Sphingosine-1-phosphate lyase insufficiency syndrome (SPLIS) is a rare metabolic disorder caused by inactivating mutations in sphingosine-1-phosphate lyase 1 (SGPL1), which is required for the final step of sphingolipid metabolism. SPLIS features include steroid-resistant nephrotic syndrome and impairment of neurological, endocrine, and hematopoietic systems. Many affected individuals die within the first 2 years. No targeted therapy for SPLIS is available. We hypothesized that SGPL1 gene replacement would address the root cause of SPLIS, thereby serving as a universal treatment for the condition. As proof of concept, we evaluated the efficacy of adeno-associated virus 9-mediated transfer of human SGPL1 (AAV-SPL) given to newborn Sgpl1-KO mice that model SPLIS and die in the first weeks of life. Treatment dramatically prolonged survival and prevented nephrosis, neurodevelopmental delay, anemia, and hypercholesterolemia. STAT3 pathway activation and elevated proinflammatory and profibrogenic cytokines observed in KO kidneys were attenuated by treatment. Plasma and tissue sphingolipids were reduced in treated compared with untreated KO pups. SGPL1 expression and activity were measurable for at least 40 weeks. In summary, early AAV-SPL treatment prevents nephrosis, lipidosis, and neurological impairment in a mouse model of SPLIS. Our results suggest that SGPL1 gene replacement holds promise as a durable and universal targeted treatment for SPLIS.

Postsynaptic Y654 dephosphorylation of ß-catenin modulates presynaptic vesicle turnover through increased n-cadherin-mediated transsynaptic signaling.

Synaptic adhesion molecules, which coordinately control structural and functional changes at both sides of synapses, are important for synaptogenesis and synaptic plasticity. Because they physically form homophilic or heterophilic adhesions across synaptic junctions, these molecules can initiate transsynaptic communication in both anterograde and retrograde directions. Using optical imaging approaches, we investigated whether an increase in postsynaptic N-cadherin could correspondingly alter the function of connected presynaptic terminals. Postsynaptic expression of ß-catenin Y654F, a phosphorylation-defective form with enhanced binding to N-cadherin, is sufficient to increase postsynaptic surface levels of N-cadherin and consequently promote presynaptic reorganizations. Such reorganizations include increases in the densities of the synaptic vesicle protein, Synaptotagmin 1 and the active zone scaffold protein, Bassoon, the number of active boutons and the size of the total recycling vesicle pool. In contrast, synaptic vesicle turnover is significantly impaired, preventing the exchange of synaptic vesicles with adjacent boutons. Together, N-cadherin-mediated retrograde signaling, governed by phosphoregulation of postsynaptic ß-catenin Y654, coordinately modulates presynaptic vesicle dynamics to enhance synaptic communication in mature neurons. © 2016 Wiley Periodicals, Inc. Develop Neurobiol 77: 61-74, 2017.

N-Cadherin Regulates Cell Migration Through a Rab5-Dependent Temporal Control of Macropinocytosis.

Macropinocytosis is a clathrin-independent endocytic pathway implicated in fluid uptake, pathogen invasion and cell migration. During collective cell migration, macropinocytosis occurs primarily at membrane ruffles arising from the leading edges of migrating cells. We report here that N-cadherin (Ncad) regulates the tempo of macropinocytosis and thereby influences wound-induced collective cell migration. Using live-cell and super-resolution imaging techniques, we observed that Ncad formed clusters at the membrane ruffles and macropinosomes. De-clustering of Ncad by an interfering antibody impaired the recruitment of Rab5-an early endosomal marker-to the macropinosomes. Moreover, we demonstrated that Ncad interacts with Rab5, and laser ablation of Ncad caused Rab5 to dissociate from the macropinosomes. Although Rab5 detached from macropinosomes upon the de-clustering of Ncad, the recruitment of late endosomal marker Rab7 occurred earlier. Consequently, both centripetal trafficking of macropinosomes and collective migration were accelerated due to de-clustering of Ncad. Thus, our results suggest that Ncad is involved in the maturation of macropinocytosis through Rab5 recruitment, linking macropinocytosis and cell migration through a novel function of Ncad.

Autographa californica multiple nucleopolyhedrovirus LEF-2 is a capsid protein required for amplification but not initiation of viral DNA replication.

The late expression factor 2 gene (lef-2) of baculovirus Autographa californica multiple nucleopolyhedrovirus (AcMNPV) has been identified as one of the factors essential for origin-dependent DNA replication in transient expression assays and has been shown to be involved in late/very late gene expression. To study the function of lef-2 in the life cycle of AcMNPV, lef-2 knockout and repair bacmids were generated by homologous recombination in Escherichia coli. Growth curve analysis showed that lef-2 was essential for virus production. Interestingly, a DNA replication assay indicated that lef-2 is not required for the initiation of viral DNA replication and that, rather, it is required for the amplification of DNA replication. lef-2 is also required for the expression of late and very late genes, as the expression of these genes was abolished by lef-2 deletion. Temporal and spatial distributions of LEF-2 protein in infected cells were also analyzed, and the data showed that LEF-2 protein was localized to the virogenic stroma in the nuclei of the infected cells. Analysis of purified virus particles revealed that LEF-2 is a viral protein component of both budded and occlusion-derived virions, predominantly in the nucleocapsids of the virus particles. This observation suggests that LEF-2 may be required immediately after virus entry into host cells for efficient viral DNA replication.

A functional genomics approach to identify and characterize oxidation resistance genes.

In order to develop a more complete understanding of the genes required for resistance to oxidative DNA damage, we devised methods to identify genes that can prevent or repair oxidative DNA damage. These methods use the oxidative mutator phenotype of a repair deficient E. coli strain to measure the antimutator effect resulting from the expression of human cDNAs. The method can be adapted to characterize the function, and to determine the active site domains, of putative antimutator genes. Since bacteria do not contain subcellular compartments, genes that function in mitochondria, the cytoplasm, or the nucleus can be identified. Methods to determine the localization of genes in their normal host organism are also described.

Identification of baculoviral factors required for the activation of enhancer-like polyhedrin upstream (pu) sequence.

Previously, we identified a novel enhancer-like element, the polyhedrin upstream (pu) sequence, in the genome of the baculovirus Autographa californica multiple nucleopolyhedrovirus (AcMNPV), which activates several early promoters. The activation requires co-infection of AcMNPV, suggesting that viral gene products are needed for pu-mediated promoter activation. DNA replication assay showed that the pu sequence did not assist in DNA replication and suggested its involvement in activated transcription from target promoters. In order to identify the viral genes responsible for pu-dependent activation of early promoters, a set of overlapping cosmid clones covering the entire 134-kb AcMNPV genome were constructed and screened. Our results identified three viral genes ie1, ie2, and pe38, which function in concert with pu to activate target promoters. In addition, these three viral factors can substitute for the entire virus for the synergistic promoter activation mediated by pu and the known baculovirus enhancer, the homologous region.

The single-strand DNA binding activity of human PC4 prevents mutagenesis and killing by oxidative DNA damage.

Human positive cofactor 4 (PC4) is a transcriptional coactivator with a highly conserved single-strand DNA (ssDNA) binding domain of unknown function. We identified PC4 as a suppressor of the oxidative mutator phenotype of the Escherichia coli fpg mutY mutant and demonstrate that this suppression requires its ssDNA binding activity. Saccharomyces cerevisiae mutants lacking their PC4 ortholog Sub1 are sensitive to hydrogen peroxide and exhibit spontaneous and peroxide-induced hypermutability. PC4 expression suppresses the peroxide sensitivity of the yeast sub1Delta mutant, suggesting that the human protein has a similar function. A role for yeast and human proteins in DNA repair is suggested by the demonstration that Sub1 acts in a peroxide resistance pathway involving Rad2 and by the physical interaction of PC4 with the human Rad2 homolog XPG. We show that XPG recruits PC4 to a bubble-containing DNA substrate with a resulting displacement of XPG and formation of a PC4-DNA complex. We discuss the possible requirement for PC4 in either global or transcription-coupled repair of oxidative DNA damage to mediate the release of XPG bound to its substrate.