BackgroundMagnetic resonance imaging (MRI) of the knee is the preferred method for diagnosing knee injuries. However, interpretation of knee MRI is time-intensive and subject to diagnostic error and variability. An automated system for interpreting knee MRI could prioritize high-risk patients and assist clinicians in making diagnoses. Deep learning methods, in being able to automatically learn layers of features, are well suited for modeling the complex relationships between medical images and their interpretations. In this study we developed a deep learning model for detecting general abnormalities and specific diagnoses (anterior cruciate ligament [ACL] tears and meniscal tears) on knee MRI exams. We then measured the effect of providing the model’s predictions to clinical experts during interpretation.Methods and findingsOur dataset consisted of 1,370 knee MRI exams performed at Stanford University Medical Center between January 1, 2001, and December 31, 2012 (mean age 38.0 years; 569 [41.5%] female patients). The majority vote of 3 musculoskeletal radiologists established reference standard labels on an internal validation set of 120 exams. We developed MRNet, a convolutional neural network for classifying MRI series and combined predictions from 3 series per exam using logistic regression. In detecting abnormalities, ACL tears, and meniscal tears, this model achieved area under the receiver operating characteristic curve (AUC) values of 0.937 (95% CI 0.895, 0.980), 0.965 (95% CI 0.938, 0.993), and 0.847 (95% CI 0.780, 0.914), respectively, on the internal validation set. We also obtained a public dataset of 917 exams with sagittal T1-weighted series and labels for ACL injury from Clinical Hospital Centre Rijeka, Croatia. On the external validation set of 183 exams, the MRNet trained on Stanford sagittal T2-weighted series achieved an AUC of 0.824 (95% CI 0.757, 0.892) in the detection of ACL injuries with no additional training, while an MRNet trained on the rest of the external data achieved an AUC of 0.911 (95% CI 0.864, 0.958). We additionally measured the specificity, sensitivity, and accuracy of 9 clinical experts (7 board-certified general radiologists and 2 orthopedic surgeons) on the internal validation set both with and without model assistance. Using a 2-sided Pearson’s chi-squared test with adjustment for multiple comparisons, we found no significant differences between the performance of the model and that of unassisted general radiologists in detecting abnormalities. General radiologists achieved significantly higher sensitivity in detecting ACL tears (p-value = 0.002; q-value = 0.019) and significantly higher specificity in detecting meniscal tears (p-value = 0.003; q-value = 0.019). Using a 1-tailed t test on the change in performance metrics, we found that providing model predictions significantly increased clinical experts’ specificity in identifying ACL tears (p-value < 0.001; q-value = 0.006). The primary limitations of our study include lack of surgical ground truth and the small size of ...

The cyclin D1 gene encodes the regulatory subunit of a holoenzyme that phosphorylates and inactivates the pRB tumor suppressor protein. Cyclin D1 is overexpressed in 20 -30% of human breast tumors and is induced both by oncogenes including those for Ras, Neu, and Src, and by the ␤-catenin/lymphoid enhancer factor (LEF)/T cell factor (TCF) pathway. The ankyrin repeat containing serine-threonine protein kinase, integrinlinked kinase (ILK), binds to the cytoplasmic domain of ␤ 1 and ␤ 3 integrin subunits and promotes anchorageindependent growth. We show here that ILK overexpression elevates cyclin D1 protein levels and directly induces the cyclin D1 gene in mammary epithelial cells. ILK activation of the cyclin D1 promoter was abolished by point mutation of a cAMP-responsive element-binding protein (CREB)/ATF-2 binding site at nucleotide ؊54 in the cyclin D1 promoter, and by overexpression of either glycogen synthase kinase-3␤ (GSK-3␤) or dominant negative mutants of CREB or ATF-2. Inhibition of the PI 3-kinase and AKT/protein kinase B, but not of the p38, ERK, or JNK signaling pathways, reduced ILK induction of cyclin D1 expression. ILK induced CREB transactivation and CREB binding to the cyclin D1 promoter CRE. Wnt-1 overexpression in mammary epithelial cells induced cyclin D1 mRNA and targeted overexpression of Wnt-1 in the mammary gland of transgenic mice increased both ILK activity and cyclin D1 levels. We conclude that the cyclin D1 gene is regulated by the Wnt-1 and ILK signaling pathways and that ILK induction of cyclin D1 involves the CREB signaling pathway in mammary epithelial cells.The cyclin D1 gene encodes a regulatory subunit of a serinethreonine kinase that phosphorylates and inactivates the tumor suppressor pRB (1). The abundance of cyclin D1 was shown to be rate-limiting in cellular proliferation induced by diverse signaling pathways in fibroblasts and breast epithelial cells, including MCF7 cells (2, 3). Homozygous deletion of the cyclin D1 gene in mice results in defects in mammary gland development (4, 5) and in serum-induced cellular proliferation (6). The abundance of cyclin D1 is increased in more than 30% of human breast tumors, and overexpression of cyclin D1 under control of the MMTV 1 promoter in transgenic mice induces mammary adenocarcinoma (7). The majority of breast cancer cell lines and mammary tumors induced by transgenic overexpression of either pp60 v-src or ErbB-2 oncogenes overexpress cyclin D1, suggesting that the induction of cyclin D1 may play an important role in mammary tumorigenesis (8). The cyclin D1 gene is activated by mitogenic stimuli induced by G-protein-

Transcriptional repression mechanisms are important during differentiation of multipotential hematopoietic progenitors, where they are thought to regulate lineage commitment and to extinguish alternative differentiation programs. PU.1 and GATA‐1 are two critical hematopoietic transcription factors that physically interact and mutually antagonize each other's transcriptional activity and ability to promote myeloid and erythroid differentiation, respectively. We find that PU.1 inhibits the erythroid program by binding to GATA‐1 on its target genes and organizing a complex of proteins that creates a repressive chromatin structure containing lysine‐9 methylated H3 histones and heterochromatin protein 1. Although these features are thought to be stable aspects of repressed chromatin, we find that silencing of PU.1 expression leads to removal of the repression complex, loss of the repressive chromatin marks and reactivation of the erythroid program. This process involves incorporation of the replacement histone variant H3.3 into nucleosomes. Repression of one transcription factor bound to DNA by another transcription factor not on the DNA represents a new mechanism for downregulating an alternative gene expression program during lineage commitment of multipotential hematopoietic progenitors.

No abstract

EDUCATIONAL OBJECTIVES As a result of reading this article, physicians should be able to: 1. Identify at-risk populations for giant cell tumor of bone. 2. Recognize the biology that drives giant cell tumor of bone. 3. Describe modern surgical and adjuvant techniques to effectively treat giant cell tumor of bone. 4. Recognize the complications associated with radiation therapy, poor resection, and adjuvant treatments. Giant cell tumor of bone (GCT) is a benign, locally aggressive bone tumor. Giant cell tumor of bone primarily affects the young adult patient population. The natural history of GCT is progressive bone destruction leading to joint deformity and disability. Surgery is the primary mode of treatment, but GCT has a tendency to recur locally despite a range of adjuvant surgical options. Pulmonary metastasis has been described. However, systemic spread of GCT rarely becomes progressive, leading to death. This review presents the clinicopathologic features of GCT and a historical perspective that highlights the current rationale and controversies regarding the treatment of GCT.