A neural network model for detection and classification of lumbar spinal stenosis on MRI.

OBJECTIVES: To develop a three-stage convolutional neural network (CNN) approach to segment anatomical structures, classify the presence of lumbar spinal stenosis (LSS) for all 3 stenosis types: central, lateral recess and foraminal and assess its severity on spine MRI and to demonstrate its efficacy as an accurate and consistent diagnostic tool. METHODS: The three-stage model was trained on 1635 annotated lumbar spine MRI studies consisting of T2-weighted sagittal and axial planes at each vertebral level. Accuracy of the model was evaluated on an external validation set of 150 MRI studies graded on a scale of absent, mild, moderate or severe by a panel of 7 radiologists. The reference standard for all types was determined by majority voting and in case of disagreement, adjudicated by an external radiologist. The radiologists' diagnoses were then compared to the diagnoses of the model. RESULTS: The model showed comparable performance to the radiologist average both in terms of the determination of presence/absence of LSS as well as severity classification, for all 3 stenosis types. In the case of central canal stenosis, the sensitivity, specificity and AUROC of the CNN were (0.971, 0.864, 0.963) for binary (presence/absence) classification compared to the radiologist average of (0.786, 0.899, 0.842). For lateral recess stenosis, the sensitivity, specificity and AUROC of the CNN were (0.853, 0.787, 0.907) compared to the radiologist average of (0.713, 0.898, 805). For foraminal stenosis, the sensitivity, specificity and AUROC of the CNN were (0.942, 0.844, 0.950) compared to the radiologist average of (0.879, 0.877, 0.878). Multi-class severity classifications showed similarly comparable statistics. CONCLUSIONS: The CNN showed comparable performance to radiologist subspecialists for the detection and classification of LSS. The integration of neural network models in the detection of LSS could bring higher accuracy, efficiency, consistency, and post-hoc interpretability in diagnostic practices.

FTY720-enhanced T cell homing is dependent on CCR2, CCR5, CCR7, and CXCR4: evidence for distinct chemokine compartments.

FTY720 stimulates CCR7-driven T cell homing to peripheral lymph nodes (LN) by direct activation of sphingosine 1-phosphate receptors, along with the participation of multidrug transporters, 5-lipoxygenase, and G protein-coupled receptors for chemokines. In this study, we demonstrate that FTY720 also directly stimulates in vitro T cell chemotaxis to CCR2-CCL2, but not to a variety of other chemokines, including CCR5-CCL3/4/5 and CXCR4-CXCL12. FTY720 influences CCR2-CCL2-driven migration through activation of the multidrug transporters, Abcb1 and Abcc1, and through 5-lipoxygenase activity. In vivo administration of FTY720 induces chemokine-dependent migration of T cells in the thymus, peripheral blood, LN, and spleen. The CCR7 and CCR2 chemokine ligands are required for both T cell sequestration in LN and thymic T cell egress following FTY720 administration. Furthermore, FTY720 administration uncovers a requirement for CXCR4 ligands for LN homing, but not for thymic egress, and CCR5 for thymic egress, but not LN homing. FTY720-driven splenic and peripheral blood T cell egress are both independent of CCR2, CCR5, CCR7, or CXCR4. These results indicate that FTY720- and sphingosine 1-phosphate receptor-stimulated T cell migration are dependent on the restricted usage of chemokine receptor-ligand pairs within discrete anatomic compartments.

FTY720 stimulates multidrug transporter- and cysteinyl leukotriene-dependent T cell chemotaxis to lymph nodes.

FTY720 is a sphingosine-derived immunosuppressant. Phosphorylated FTY720 promotes T cell homing from spleen and peripheral blood to LNs by acting as an agonist for sphingosine-1-phosphate (S1P) receptors. Here we demonstrate that FTY720 enhances the activity of the sphingosine transporter Abcb1 (Mdr1) and the leukotriene C(4) transporter Abcc1 (Mrp1). Both transporters must be active for FTY720-mediated T cell migration and LN homing. Migration and homing driven by FTY720, phosphorylated FTY720, or S1P also require 5-lipoxygenase-mediated synthesis of cysteinyl leukotrienes and their efflux from the cell. FTY720-mediated LN homing events further downstream are dependent on CCL19, CCL21, VLA-4alpha, and CD44. Use of T cells deficient in 5-lipoxygenase, Abcb1, and Abcc1, and comparison of the effects of FTY720 with those of S1P, suggest a model of sequential engagement of Abcb1, SP1 receptors, 5-lipoxygenase, and Abcc1 to enhance T cell migration and homing.

CD2 is a dominant target for allogeneic responses.

CD2 and 2B4 (CD244) are members of the immunoglobulin gene superfamily and are both ligands for another family member, CD48. CD2 is widely distributed on T, NK, and B cells and some antigen-presenting cells, while 2B4 is expressed on NK and some T cells and monocytes and is known to participate in NK cytotoxicity. Since indefinite allograft survival could be obtained by a combination of anti-CD48 plus anti-CD2 mAb administration, it was important to determine the role of 2B4 blockade in allograft rejection. MAbs directed against CD2, CD48, or 2B4 were administered singly or in pairs to cardiac allograft recipients. The experiments show that only anti-CD2 plus anti-CD48 mAbs result in indefinite allograft survival, while anti-CD2 plus anti-2B4 mAbs substantially prolong graft survival, and anti-CD48 plus anti-2B4 mAbs were no better than each mAb alone. The effect of these mAbs on anti-CD3 mAb and alloantigen-driven proliferation and IFN-gamma production were also assessed. In general, anti-CD2 inhibited both anti-CD3 mAb and alloantigen-driven responses, while anti-CD48 inhibited only anti-CD3 mAb but not alloantigen-driven proliferative and cytokine responses. Anti-2B4 mAbs were generally ineffective alone. Combinations of mAbs were more effective than single mAbs only in alloantigen-driven proliferation, commensurate with allograft survival results. Using CD2-/- and CD48-/- T cells and antigen-presenting cells, we also demonstrate that these inhibitory mAbs act primarily by blocking intercellular interactions, rather than directly delivering negative signals to T cells. These results suggest that, unlike CD2, 2B4 is not a potent regulatory molecule or ligand for CD48 in the response to alloantigen. Blocking the 2B4-CD48 receptor-ligand pair does not inhibit T-cell responses and alloreactivity to the same degree as CD2-CD48 blockade.

L-selectin-dependent lymphoid occupancy is required to induce alloantigen-specific tolerance.

Maneuvers that interfere with signals 1, 2, 3, or Ag processing can result in indefinite allograft survival. However, they are not applicable to all tissues, strains, or species, suggesting that there are additional levels of immune regulation. We hypothesized that secondary lymphoid organs are important for interactions among lymphocytes, alloantigen, and immunosuppressants that lead to tolerance. To explore this, cardiac allografts were performed with a tolerogenic immunosuppressive regimen. Concurrent administration of anti-L-selectin (CD62L) Ab, which prevents lymph node homing, prevents indefinite allograft survival and tolerance. Anti-CD62L Ab is not costimulatory, and Fab and F(ab')(2) anti-CD62L have similar activities. Flow cytometry and histologic examination show that Ab shifts T cells away from lymph nodes and into spleen, peripheral blood, and graft. Tolerance is not induced in CD62L(-/-) mice, and adoptive transfer of CD62L(-/-), but not CD62L(+/+), T cells prevents tolerization in wild-type recipients. FTY720, an immunosuppressant that promotes chemokine-dependent, but CD62L-independent, lymph node homing, reverses the Ab effect. Blockade of other homing receptors also prevents tolerization. These results indicate that T lymphocytes use CD62L-dependent migration for alloantigen-specific tolerance, and suggest that lymph nodes or other lymphoid tissues are an important site for peripheral tolerization to alloantigen.