Citations

Background: Prostate cancer is the second most common new cancer diagnosis in the United States. It is usually slow-growing, and when it is low-grade and confined to the prostate gland, it can be treated either conservatively (through active surveillance) or with surgery. However, if the cancer has spread beyond the prostate, such as to the lymph nodes, then that indicates a more aggressive cancer, and surgery may not be adequate. Methods: The challenge is that it is often difficult for radiologists reading prostate-specific imaging such as magnetic resonance images (MRIs) to differentiate malignant lymph nodes from non-malignant ones. An emerging field is the development of artificial intelligence (AI) models, including machine learning and deep learning, for medical imaging to assist in diagnostic tasks. Earlier research focused on implementing texture algorithms to extract imaging features used in classification models. More recently, researchers began studying the use of deep learning for both stand-alone feature extraction and end-to-end classification tasks. In order to tackle the challenges inherent in small datasets, this study was designed as a scalable hybrid framework utilizing pre-trained ResNet-18, a deep learning model, to extract features that were subsequently fed into a machine learning classifier to automatically identify malignant lymph nodes in patients with prostate cancer. For comparison, two texture algorithms were implemented, namely the gray-level co-occurrence matrix (GLCM) and Gabor. Results: Using an institutional prostate lymph node dataset (42 positives, 84 negatives), the proposed framework achieved an accuracy of 76.19%, a sensitivity of 79.76%, and a specificity of 69.05%. Using GLCM features, the classification achieved an accuracy of 61.90%, a sensitivity of 74.07%, and a specificity of 42.86%. Using Gabor features, the classification achieved an accuracy of 65.08%, a sensitivity of 73.47%, and a specificity of 52.50%. Conclusions: Our results demonstrate that a hybrid approach, i.e., using a pre-trainined deep learning model for feature extraction, followed by a machine learning classifier, is a viable solution. This hybrid approach is especially useful in medical-imaging-based applications with small datasets.

BACKGROUND: A hypothetical concern has been raised that sacubitril/valsartan might cause cognitive impairment because neprilysin is one of several enzymes degrading amyloid-β peptides in the brain, some of which are neurotoxic and linked to Alzheimer-type dementia. To address this, we examined the effect of sacubitril/valsartan compared with valsartan on cognitive function in patients with heart failure with preserved ejection fraction in a prespecified substudy of PARAGON-HF (Prospective Comparison of Angiotensin Receptor Neprilysin Inhibitor With Angiotensin Receptor Blocker Global Outcomes in Heart Failure With Preserved Ejection Fraction). METHODS: In PARAGON-HF, serial assessment of cognitive function was conducted in a subset of patients with the Mini-Mental State Examination (MMSE; score range, 0–30, with lower scores reflecting worse cognitive function). The prespecified primary analysis of this substudy was the change from baseline in MMSE score at 96 weeks. Other post hoc analyses included cognitive decline (fall in MMSE score of ≥3 points), cognitive impairment (MMSE score <24), or the occurrence of dementia-related adverse events. RESULTS: Among 2895 patients included in the MMSE substudy with baseline MMSE score measured, 1453 patients were assigned to sacubitril/valsartan and 1442 to valsartan. Their mean age was 73 years, and the median follow-up was 32 months. The mean±SD MMSE score at randomization was 27.4±3.0 in the sacubitril/valsartan group, with 10% having an MMSE score <24; the corresponding numbers were nearly identical in the valsartan group. The mean change from baseline to 96 weeks in the sacubitril/valsartan group was −0.05 (SE, 0.07); the corresponding change in the valsartan group was −0.04 (0.07). The mean between-treatment difference at week 96 was −0.01 (95% CI, −0.20 to 0.19; P =0.95). Analyses of a ≥3-point decline in MMSE, decrease to a score <24, dementia-related adverse events, and combinations of these showed no difference between sacubitril/valsartan and valsartan. No difference was found in the subgroup of patients tested for apolipoprotein E ε4 allele genotype. CONCLUSIONS: Patients with heart failure with preserved ejection fraction in PARAGON-HF had relatively low baseline MMSE scores. Cognitive change, measured by MMSE, did not differ between treatment with sacubitril/valsartan and treatment with valsartan in patients with heart failure with preserved ejection fraction. REGISTRATION: URL: https://www.clinicaltrials.gov ; Unique identifier: NCT01920711.

Abstract Background and Aims Perianal fistuliing Crohn’s disease [PFCD] is an aggressive phenotype of Crohn’s disease defined by frequent relapses and disabling symptoms. A novel consensus classification system was recently outlined by the TOpCLASS consortium, which seeks to unify disease severity with patient-centred goals but has not yet been validated. We aimed to apply this to a real-world cohort and to identify factors that predict transition between classes over time. Methods We identified all patients with PFCD and at least one baseline and one follow-up pelvic MRI [pMRI]. TOpCLASS classification, disease characteristics, and imaging indices were collected retrospectively at time periods corresponding with respective MRIs. Results We identified 100 patients with PFCD, of whom 96 were assigned TOpCLASS Classes 1–2c at baseline. Most patients [78.1%] started in Class 2b, but changes in classification were observed in 52.1% of all patients. Male sex [72.0%, 46.6%, 40.0%, p = 0.03] and prior perianal surgery [52.0% vs 44.6% vs 40.0%, p = 0.02] were more frequently observed in those with improved class compared to unchanged and worsened class. Baseline pMRI indices were not associated with changes in classification; however, greater improvements in mVAI, MODIFI-CD, and PEMPAC were seen among those who improved. Linear mixed effect modelling identified only male sex [-0.31, 95% CI -0.60 to -0.02] with improvement in class. Conclusion The TOpCLASS classification highlights the dynamic nature of PFCD over time. However, our ability to predict transitions between classes remains limited and requires prospective assessment. Improvement in MRI index scores over time was associated with a transition to lower TOpCLASS classification.

Abstract Inactivating mutations of genes encoding the cohesin complex are common in a wide range of human cancers. STAG2 is the most commonly mutated cohesin complex subunit. Here we report the impact of stable correction of endogenous, naturally occurring STAG2 mutations on gene expression, 3D genome organization, chromatin loops, and Polycomb signaling in glioblastoma multiforme (GBM). In two GBM cell lines, correction of their STAG2 mutations significantly altered the expression of ~10% of all expressed genes. Virtually all the most highly regulated genes were negatively regulated by STAG2 (i.e., expressed higher in STAG2-mutant cells than in STAG2-corrected cells), and one of them – HEPH – was regulated by STAG2 in uncultured GBM tumors as well. While STAG2 correction had little effect on large scale features of 3D genome organization (A/B compartments, TADs), STAG2 correction did alter thousands of individual chromatin loops, some of which controlled the expression of adjacent genes. Loops specific to STAG2-mutant cells, which were regulated by STAG1-containing cohesin complexes, were very large, supporting prior findings that STAG1-containing cohesin complexes have greater loop extrusion processivity than STAG2-containing cohesin complexes and suggesting that long loops may be a general feature of STAG2-mutant cancers. Finally, STAG2-regulated chromatin loops were enriched for the H3K27me3 Polycomb epigenetic mark, revealing that inactivation of STAG2 can activate Polycomb signaling in GBM, identifying a potential target for therapeutic intervention in STAG2-mutant tumors. Together, these findings illuminate the landscape of STAG2-regulated genes, A/B compartments, chromatin loops, and pathways in GBM, providing important clues into the largely still unknown mechanism of STAG2 tumor suppression.

Surgical treatment for Legg–Calve–Perthes disease (LCPD) is recommended for older children with moderate to severe disease. We sought to determine whether double osteotomies lead to improved radiologic outcomes compared to reported non-operative outcomes. Patients older than 6 years of age diagnosed with LCPD lateral pillar B or C who were treated with pelvic and femoral osteotomies were included. Radiologic outcomes and leg-length discrepancies were assessed using the Stulberg classification and were compared with the current literature. Fifteen hips in fourteen patients were treated with double osteotomy for LCPD, and seven had lateral pillar C disease (47%). The mean age at surgery was 8.6 years (range, 7.2–10.4) and the mean age at follow-up was 20.2 years (range, 14.2–35.6). At a mean 11.6-year follow-up (range: 6.3–25.2), double osteotomy resulted in 40% of patients having Stulberg I/II scores, 27% having Stulberg III scores, and 33% having Stulberg IV/V scores. The mean leg-length discrepancy was 1.4 cm in lateral pillar C patients compared to 0.8 cm in lateral pillar B patients. Four patients underwent additional surgeries, including two who required total hip arthroplasty. Double osteotomy as an alternative surgical procedure for the treatment of LCPD did not show improved outcomes when compared to historic non-operative cohorts.

Background For acute proximal intracranial artery occlusions, contact aspiration may be more effective than stent‐retriever for first‐line reperfusion therapy. Due to the lack of data regarding medium vessel occlusion thrombectomy, we evaluated outcomes according to first‐line technique in a large, multicenter registry. Methods Imaging, procedural, and clinical outcomes of patients with acute proximal medium vessel occlusions (M2, A1, or P1) or distal medium vessel occlusions (M3, A2, P2, or further) treated at 37 sites in 10 countries were analyzed according to first‐line endovascular technique (stent‐retriever versus aspiration). Multivariable logistic regression and propensity‐score matching were used to estimate the odds of the primary outcome, expanded Thrombolysis in Cerebral Infarction score of 2b–3 (“successful recanalization”), as well as secondary outcomes (first‐pass effect, expanded Thrombolysis in Cerebral Infarction 2c‐3, intracerebral hemorrhage, and 90‐day modified Rankin scale, 90‐day mortality) between treatment groups. Results Of the 440 included patients (44.5% stent‐retriever versus 55.5% aspiration), those treated with stent‐retriever had lower baseline Alberta Stroke Program Early Computed Tomography Scale scores (median 8 versus 9; P <0.01), higher National Institutes of Health Stroke Scale scores (median 13 versus 11; P =0.02), and nonsignificantly fewer medium‐distal occlusions (M3, A2, P2, or other: 17.4% versus 23.8%; P =0.10). Use of a stent‐retriever was associated with 15% lower odds of successful recanalization (odds ratio [OR], 0.85; [95% CI 0.74–0.98]; P =0.02), but this was not significant after multivariable adjustment in the total cohort (adjusted OR, 0.88; [95% CI 0.72–1.09]; P =0.24), or in the propensity‐score matched cohort (n=105 in each group) (adjusted OR, 0.94; [95% CI 0.75–1.18]; P =0.60). There was no significant association between technique and secondary outcomes in the propensity‐score matched adjusted models. Conclusion In this large, diverse, multinational medium vessel occlusion cohort, we found no significant difference in imaging or clinical outcomes with aspiration versus stent‐retriever thrombectomy.

AbstractCheckpoint inhibitor (CPI) therapy with anti–PD-1 antibodies has been associated with mixed outcomes in small cohorts of patients with relapsed aggressive B-cell lymphomas after CAR-T failure. To define CPI therapy efficacy more definitively in this population, we retrospectively evaluated clinical outcomes in a large cohort of 96 patients with aggressive B-cell lymphomas receiving CPI therapy after CAR-T failure across 15 US academic centers. Most patients (53%) had diffuse large B-cell lymphoma, were treated with axicabtagene ciloleucel (53%), relapsed early (≤180 days) after CAR-T (83%), and received pembrolizumab (49%) or nivolumab (43%). CPI therapy was associated with an overall response rate of 19% and a complete response rate of 10%. Median duration of response was 221 days. Median progression-free survival (PFS) and overall survival (OS) were 54 and 159 days, respectively. Outcomes to CPI therapy were significantly improved in patients with primary mediastinal B-cell lymphoma. PFS (128 vs 51 days) and OS (387 vs 131 days) were significantly longer in patients with late (&gt;180 days) vs early (≤180 days) relapse after CAR-T. Grade ≥3 adverse events occurred in 19% of patients treated with CPI. Most patients (83%) died, commonly because of progressive disease. Only 5% had durable responses to CPI therapy. In the largest cohort of patients with aggressive B-cell lymphoma treated with CPI therapy after CAR-T relapse, our results reveal poor outcomes, particularly among those relapsing early after CAR-T. In conclusion, CPI therapy is not an effective salvage strategy for most patients after CAR-T, where alternative approaches are needed to improve post–CAR-T outcomes.

Accurate clinical staging of bladder cancer aids in optimizing the process of clinical decision-making, thereby tailoring the effective treatment and management of patients. While several radiomics approaches have been developed to facilitate the process of clinical diagnosis and staging of bladder cancer using grayscale computed tomography (CT) scans, the performances of these models have been low, with little validation and no clear consensus on specific imaging signatures. We propose a hybrid framework comprising pre-trained deep neural networks for feature extraction, in combination with statistical machine learning techniques for classification, which is capable of performing the following classification tasks: (1) bladder cancer tissue vs. normal tissue, (2) muscle-invasive bladder cancer (MIBC) vs. non-muscle-invasive bladder cancer (NMIBC), and (3) post-treatment changes (PTC) vs. MIBC.