| 1. |
- Rudà, Roberta, et al.
(author)
-
Eano-Euracan-Sno Guidelines on Circumscribed Astrocytic GLIOMAS, Glioneuronal and Neuronaltumors.
- 2022
-
In: Neuro-oncology. - : Oxford University Press (OUP). - 1523-5866 .- 1522-8517. ; 24:12, s. 2015-2034
-
Research review (peer-reviewed)abstract
- In the new WHO 2021 Classification of CNS Tumors the chapter "Circumscribed astrocytic gliomas, glioneuronal and neuronal tumors" encompasses several different rare tumor entities, which occur more frequently in children, adolescents and young adults. The Task Force has reviewed the evidence of diagnostic and therapeutic interventions, which is low particularly for adult patients, and draw recommendations accordingly. Tumor diagnosis, based on WHO 2021, is primarily performed using conventional histological techniques; however, molecular workup is important for differential diagnosis, in particular DNA methylation profiling for the definitive classification of histologically unresolved cases. Molecular factors are increasingly of prognostic and predictive importance. MRI finding are non specific, but for some tumors are characteristic and suggestive. Gross total resection, when feasible, is the most important treatment in terms of prolonging survival and achieving long-term seizure control. Conformal radiotherapy should be considered in grade 3 and incompletely resected grade 2 tumors. In recurrent tumors reoperation and radiotherapy, including stereotactic radiotherapy, can be useful. Targeted therapies may be used in selected patients: BRAF and MEK inhibitors in pilocytic astrocytomas, pleomorphic xanthoastrocytomas, and gangliogliomas when BRAF altered, and mTOR inhibitor everolimus in subependymal giant cells astrocytomas .Sequencing to identify molecular targets is advocated for diagnostic clarification and to direct potential targeted therapies.
|
|
| 2. |
- Stokes, Chris R., et al.
(author)
-
On the reconstruction of palaeo-ice sheets : Recent advances and future challenges
- 2015
-
In: Quaternary Science Reviews. - : Elsevier BV. - 0277-3791 .- 1873-457X. ; 125, s. 15-49
-
Research review (peer-reviewed)abstract
- Reconstructing the growth and decay of palaeo-ice sheets is critical to understanding mechanisms of global climate change and associated sea-level fluctuations in the past, present and future. The significance of palaeo-ice sheets is further underlined by the broad range of disciplines concerned with reconstructing their behaviour, many of which have undergone a rapid expansion since the 1980s. In particular, there has been a major increase in the size and qualitative diversity of empirical data used to reconstruct and date ice sheets, and major improvements in our ability to simulate their dynamics in numerical ice sheet models. These developments have made it increasingly necessary to forge interdisciplinary links between sub-disciplines and to link numerical modelling with observations and dating of proxy records. The aim of this paper is to evaluate recent developments in the methods used to reconstruct ice sheets and outline some key challenges that remain, with an emphasis on how future work might integrate terrestrial and marine evidence together with numerical modelling. Our focus is on pan-ice sheet reconstructions of the last deglaciation, but regional case studies are used to illustrate methodological achievements, challenges and opportunities. Whilst various disciplines have made important progress in our understanding of ice-sheet dynamics, it is clear that data-model integration remains under-used, and that uncertainties remain poorly quantified in both empirically-based and numerical ice-Sheet reconstructions. The representation of past climate will continue to be the largest source of uncertainty for numerical modelling. As such, palaeo-observations are critical to constrain and validate modelling. State-of-the-art numerical models will continue to improve both in model resolution and in the breadth of inclusion of relevant processes, thereby enabling more accurate and more direct comparison with the increasing range of palaeo-observations. Thus, the capability is developing to use all relevant palaeo-records to more strongly constrain deglacial (and to a lesser extent pre-LGM) ice sheet evolution. In working towards that goal, the accurate representation of uncertainties is required for both constraint data and model outputs. Close cooperation between modelling and data-gathering communities is essential to ensure this capability is realised and continues to progress.
|
|
| 3. |
- Umuroglu, Yaman, et al.
(author)
-
Optimizing Bit-Serial Matrix Multiplication for Reconfigurable Computing
- 2019
-
In: ACM Transactions on Reconfigurable Technology and Systems. - : ASSOC COMPUTING MACHINERY. - 1936-7406 .- 1936-7414. ; 12:3
-
Journal article (peer-reviewed)abstract
- Matrix-matrix multiplication is a key computational kernel for numerous applications in science and engineering, with ample parallelism and data locality that lends itself well to high-performance implementations. Many matrix multiplication-dependent applications can use reduced-precision integer or fixed-point representations to increase their performance and energy efficiency while still offering adequate quality of results. However, precision requirements may vary between different application phases or depend on input data, rendering constant-precision solutions ineffective. BISMO, a vectorized bit-serial matrix multiplication overlay for reconfigurable computing, previously utilized the excellent binary-operation performance of FPGAs to offer a matrix multiplication performance that scales with required precision and parallelism. We show how BISMO can be scaled up on Xilinx FPGAs using an arithmetic architecture that better utilizes six-input LUTs. The improved BISMO achieves a peak performance of 15.4 binary TOPS on the Ultra96 board with a Xilinx UltraScale+ MPSoC.
|
|
