Citations

AbstractBackgroundInaccurate manual organ delineation is one of the high‐risk failure modes in radiation treatment. Numerous automated contour quality assurance (QA) systems have been developed to assess contour acceptability; however, manual inspection of flagged cases is a time‐consuming and challenging process, and can lead to users overlooking the exact error location.PurposeOur aim is to develop and validate a contour QA system that can effectively detect and visualize subregional contour errors, both qualitatively and quantitatively.Methods/MaterialsA novel contour subregion error detection (CSED) system was developed using subregional surface distance discrepancies between manual and deep learning auto‐segmentation (DLAS) contours. A validation study was conducted using a head and neck public dataset containing 339 cases and evaluated according to knowledge‐based pass criteria derived from a clinical training dataset of 60 cases. A blind qualitative evaluation was conducted, comparing the results from the CSED system with manual labels. Subsequently, the CSED‐flagged cases were re‐examined by a radiation oncologist.ResultsThe CSED system could visualize the diverse types of subregional contour errors qualitatively and quantitatively. In the validation dataset, the CSED system resulted in true positive rates (TPR) of 0.814, 0.800, and 0.771; false positive rates (FPR) of 0.310, 0.267, and 0.298; and accuracies of 0.735, 0.759, and 0.730, for brainstem and left and right parotid contours, respectively. The CSED‐assisted manual review caught 13 brainstem, 19 left parotid, and 21 right parotid contour errors missed by conventional human review. The TPR/FPR/accuracy of the CSED‐assisted manual review improved to 0.836/0.253/0.784, 0.831/0.171/0.830, and 0.808/0.193/0.807 for each structure, respectively. Further, the time savings achieved through CSED‐assisted review improved by 75%, with the time for review taking 24.81 ± 12.84, 26.75 ± 10.41, and 28.71 ± 13.72 s for each structure, respectively.ConclusionsThe CSED system enables qualitative and quantitative detection, localization, and visualization of manual segmentation subregional errors utilizing DLAS contours as references. The use of this system has been shown to help reduce the risk of high‐risk failure modes resulting from inaccurate organ segmentation.

Clinical trials frequently include multiple end points that mature at different times. The initial report, typically based on the primary end point, may be published when key planned co-primary or secondary analyses are not yet available. Clinical Trial Updates provide an opportunity to disseminate additional results from studies, published in JCO or elsewhere, for which the primary end point has already been reported. ARST1321 was a phase II study designed to compare the near complete pathologic response rate after preoperative chemoradiation with/without pazopanib in children and adults with intermediate-/high-risk chemotherapy-sensitive body wall/extremity non-Rhabdomyosarcoma Soft Tissue Sarcoma (ClinicalTrials.gov identifier: NCT02180867 ). Enrollment was stopped early following a predetermined interim analysis that found the rate of near complete pathologic response to be significantly greater with the addition of pazopanib. As a planned secondary aim of the study, the outcome data for this cohort were analyzed. Eight-five eligible patients were randomly assigned to receive (regimen A) or not receive (regimen B) pazopanib in combination with ifosfamide and doxorubicin + preoperative radiotherapy followed by primary resection at week 13 and then further chemotherapy at week 25. As of December 31, 2021, at a median survivor follow-up of 3.3 years (range, 0.1-5.8 years), the 3-year event-free survival for all patients in the intent-to-treat analysis was 52.5% (95% CI, 34.8 to 70.2) for regimen A and 50.6% (95% CI, 32 to 69.2) for regimen B ( P = .8677, log-rank test); the 3-year overall survival was 75.7% (95% CI, 59.7 to 91.7) for regimen A and 65.4% (95% CI, 48.1 to 82.7) for regimen B ( P = .1919, log-rank test). Although the rate of near complete pathologic response was significantly greater with the addition of pazopanib, outcomes were not statistically significantly different between the two regimens.

AbstractBackgroundGene expression changes have been identified in post‐mortem brain tissue of AD cases when compared with controls, implicating various genes and pathways. Such findings provide important insights into molecular dysregulation associated with this disease and nominate therapeutic targets. Here we investigate brain gene and co‐expression network changes in a cohort of AD cases to identify associations with degree of AD neuropathology including Braak stage (Tau), Thal phase (Amyloid beta) and brain tissue levels of five AD‐related proteins (apoE, Aβ40, Aβ42, tau, and p‐tau).MethodGene expression measures were collected from temporal cortex (TCX) tissue of 477 AD cases, available through the Mayo Clinic Brain Bank, using RNA sequencing (Illumina TruSeq mRNA) performed on the HiSeq4000. Following quality control 456 cases remained for analysis. Braak and Thal were provided by an experienced neuropathologist and TCX levels of five AD‐related proteins from three fractions (buffer‐soluble, detergent‐soluble, and insoluble) were measured by ELISA (Liu et al. 2020). Weighted gene coexpression network analysis (WGCNA) R package was used to build co‐expression networks. Genes and networks were tested for association with neuropathology and AD‐related protein levels, adjusting for relevant covariates.ResultA total of 57,353 genes were profiled, of which 19,044 were reliably expressed above background levels and formed 20 co‐expression modules. 36 genes were associated with Braak stage, 10 with Thal phase and up to 9,000 with the AD‐related protein fractions. Brain levels of pTau had the greatest number of associations, with an overrepresentation of genes involved in synaptic signalling and immune response gene ontology biological processes. 16 of the coexpression modules associated with at least one trait after Bonferroni correction for number of tests.ConclusionWe have identified a number of genes and networks associated with different aspects of AD neuropathology and associated proteins, further lending support to the involvement of synaptic and immune biology in AD. However it is not clear whether observed changes cause disease or are a consequence of the neurodegenerative process. Comparison of these findings across different proteins and fractions is expected to reveal common and distinct vulnerabilities associated with amyloid and tau, and different stages of protein aggregation.

AbstractBackgroundDeep learning auto‐segmentation (DLAS) models have been adopted in the clinic; however, they suffer from performance deterioration owing to the clinical practice variability. Some commercial DLAS software provide an incremental retraining function that enables users to train a custom model using their institutional data to account for clinical practice variability.PurposeThis study was performed to evaluate and implement the commercial DLAS software with the incremental retraining function for definitive treatment of patients with prostate cancer in a multi‐user environment.MethodsCT‐based target organs and organs‐at‐risk (OAR) delineation of 215 prostate cancer patients were utilized. The performance of three commercial DLAS software built‐in models was validated with 20 patients. A retrained custom model was developed using 100 patients and evaluated on the remaining data (n = 115). Dice similarity coefficient (DSC), Hausdorff distance (HD), mean surface distance (MSD), and surface DSC (SDSC) were utilized for quantitative evaluation. A multi‐rater qualitative evaluation was blindly performed with a five‐level scale. Visual inspection was performed in consensus and non‐consensus unacceptable cases to identify the failure modes.ResultsThree commercial DLAS vendor built‐in models achieved sub‐optimal performance in 20 patients. The retrained custom model had a mean DSC of 0.82 for prostate, 0.48 for seminal vesicles (SV), and 0.92 for rectum, respectively. This represents a significant improvement over the built‐in model with DSC of 0.73, 0.37, and 0.81 for the corresponding structures. Compared to the acceptance rate of 96.5% and consensus unacceptable rate (i.e., both reviewers rated as unacceptable) of 3.5% achieved by manual contours, the custom model achieved a 91.3% acceptance rate and 8.7% consensus unacceptable rate. The failure modes of retrained custom model were attributed to the following: cystogram (n = 2), hip prosthesis (n = 2), low dose rate brachytherapy seeds (n = 2), air in endorectal balloon(n = 1), non‐iodinated spacer (n = 2), and giant bladder(n = 1).ConclusionThe commercial DLAS software with the incremental retraining function was validated and clinically adopted for prostate patients in a multi‐user environment. AI‐based auto‐delineation of the prostate and OARs is shown to achieve improved physician acceptance, overall clinical utility, and accuracy.

Abstract Purpose: Paclitaxel is a widely used anticancer therapeutic. Peripheral neuropathy is the dose-limiting toxicity and negatively impacts quality of life. Rare germline gene markers were evaluated for predicting severe taxane-induced peripheral neuropathy (TIPN) in the patients of European ancestry. In addition, the impact of Cytochrome P450 (CYP) 2C8, CYP3A4, and CYP3A5 metabolizer status on likelihood of severe TIPN was also assessed. Experimental Design: Whole-exome sequencing analyses were performed in 340 patients of European ancestry who received a standard dose and schedule of paclitaxel in the adjuvant, randomized phase III breast cancer trial, E5103. Patients who experienced grade 3–4 (n = 168) TIPN were compared to controls (n = 172) who did not experience TIPN. For the analyses, rare variants with a minor allele frequency ≤ 3% and predicted to be deleterious by protein prediction programs were retained. A gene-based, case–control analysis using SKAT was performed to identify genes that harbored an imbalance of deleterious variants associated with increased risk of severe TIPN. CYP star alleles for CYP2C8, CYP3A4, and CYP3A5 were called. An additive logistic regression model was performed to test the association of CYP2C8, CYP3A4, and CYP3A5 metabolizer status with severe TIPN. Results: Cytochrome P450 oxidoreductase (POR) was significantly associated with severe TIPN (P value = 1.8 ×10−6). Six variants were predicted to be deleterious in POR. There were no associations between CYP2C8, CYP3A4, or CYP3A5 metabolizer status with severe TIPN. Conclusions: Rare variants in POR predict an increased risk of severe TIPN in patients of European ancestry who receive paclitaxel.

Abstract (a) Expansion of therapeutic strategies against diffuse large B-cell lymphoma (DLBCL) offers new opportunities to fight aggressive non-Hodgkin's lymphoma. Therapy approaches, such as R-CHOP and stem cell transplant, and CAR-T cell therapies are now available for the treatment of DLBCL. While the role of the tumor microenvironment (TME) during the therapeutic response is now recognized, the spatial organization and the communication between the diverse factors that compose the TME remains to be fully understood. In this study, we describe a new immunogram tool to help clinical researchers measure the tumor microenvironment contribution to drug efficacy or toxicity in DLBCL context. (b) In a previous work, using a high-quality testing and innovative multimodal and integrative approach, we constructed the first Veracyte Biopharma Atlas for DLBCL lesions. Based on this knowledge and the understanding of the immune contexture of tumors, we constructed an immunogram using the scoring of clinically relevant biomarkers. Our immunogram scoring approach is composed of key features driving computational derivation of DLBCL Atlas clusters, which include immune gene signatures, immune cell phenotyping combined with spatial information, and genomic and clinical parameters. (c) Patterns that strongly associated with CAR-T treatment response were determined. In addition, promising results have been obtained in the prediction to CAR-T toxicity related to T-regulatory cells or R-CHOP response with transcriptomic signature. (d) Retrospective analysis of CAR-T clinical trial data and tumor biopsies highlighted the Veracyte Immunogram as a powerful tool to score the patient probability of response to CAR-T. This work provides the framework for identification of predictors of response to therapy, potential targets of interest or biomarkers that could be applied in future clinical trials. Citation Format: Michael D. Mattie, Regis Perbost, Sarah Turcan, Corinne Danan, Frederick L. Locke, Sattva S. Neelapu, David B. Miklos, Caron A. Jacobson, Lazaros J. Lekakis, Yi Lin, Armin Ghobadi, Jenny J. Kim, Zixing Wang, Allen Xue, Simone Filosto, Nathalie Scholler, Jerome Galon. Defining the cancer Immunogram using multimodal analysis to guide immunotherapy in diffuse large B cell lymphoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5879.

Abstract Background To determine outcomes of children with rhabdomyosarcoma (RMS) with isolated lung metastases. Methods Data were analyzed for 428 patients with metastatic RMS treated on COG protocols. Categorical variables were compared using Chi‐square or Fisher's exact tests. Event‐free survival (EFS) and overall survival (OS) were estimated using Kaplan–Meier method and compared using the log‐rank test. Results Compared with patients with other metastatic sites ( n = 373), patients with lung‐only metastases ( n = 55) were more likely to be <10 years of age, have embryonal histology (embryonal rhabdomyosarcoma), have N0 disease, and less likely to have primary extremity tumors. Lung‐only patients had significantly better survival outcomes than patients with all other sites of metastatic disease ( p < .0001) with 5‐year EFS of 48.1 versus 18.8% and 5‐year OS of 64.1 versus 26.9%. Patients with lung‐only metastases, and those with a single extrapulmonary site of metastasis, had better survival compared with patients with two or more sites of metastatic disease ( p < .0001). In patients with ERMS and lung‐only metastases, there was no significant difference in survival between patients ≥10 years and 1–9 years (5‐year EFS: 58.3 vs. 68.2%, 5‐year OS: 66.7 vs. 67.7%). Conclusions With aggressive treatment, patients with ERMS and lung‐only metastatic disease have superior EFS and OS compared with patients with other sites of metastatic disease, even when older than 10 years of age. Consideration should be given to including patients ≥10 years with ERMS and lung‐only metastases in the same group as those <10 years in future risk stratification algorithms.

Abstract Although structures of vitrified supramolecular complexes have been determined at near-atomic resolution, elucidating in situ molecular structure in living cells remains a major challenge. Here, we apply a novel but simple liquid-cell technique, developed previously for real-time imaging of the dynamics at a liquid-gas interface, to image wet biological samples. With extra scattering from the liquid phase, the transmission electron micrographs show amplitude contrast comparable to that in negatively stained samples. Single-molecule domains are resolved in the protein complex GroEL imaged in buffer solution at room temperature. Moreover, various stages of virus cell entry, which are transient events with very few structural information to date, are also captured. Morphological details are reconstructed using the technique of individual particle electron tomography. These results demonstrate that this approach can be a valuable yet cost-effective technique complementary to other microscopy techniques for addressing important biological questions at the molecular level.

AbstractBackground and ObjectivesThe impact upon wound healing of targeted molecular therapies, when incorporated into neoadjuvant therapy of soft tissue sarcoma, is largely unknown. Here, we describe wound complications following addition of pazopanib, a tyrosine kinase inhibitor (TKI), to neoadjuvant radiotherapy (RT) +/− chemotherapy for soft tissue sarcoma.MethodsWound complications were evaluated on dose‐finding and randomized arms of ARST1321, a phase II/III study incorporating neoadjuvant RT, +/− pazopanib, +/− ifosfamide/doxorubicin (ID) for sarcoma therapy.ResultsOf 85 evaluable patients, 35 (41%) experienced postoperative wound complications. Most (57%) were grade III. Randomization to pazopanib + RT + ID carried a 50% wound complication rate (17/34, with 47% grade III), compared to 22% (5/23) with ID + RT alone. In nonchemotherapy study arms, pazopanib + RT resulted in a 59% wound complication rate versus 25% for those receiving RT alone. Grade III wound complications occurred among 26% (15/58) of all patients receiving pazopanib. Wound complications occurred a median of 35 days postoperatively. Some occurred following diagnostic biopsies and at remote surgical sites.ConclusionThe addition of pazopanib to neoadjuvant chemotherapy and RT resulted in a higher wound complication rate following therapy of soft tissue sarcoma. The rate of grade III complications remained comparable to that reported in contemporary literature.