Exploiting Signal Propagation Delays to Match Task Memory Requirements in Reservoir Computing.

Recurrent neural networks (RNNs) transmit information over time through recurrent connections. In contrast, biological neural networks use many other temporal processing mechanisms. One of these mechanisms is the inter-neuron delays caused by varying axon properties. Recently, this feature was implemented in echo state networks (ESNs), a type of RNN, by assigning spatial locations to neurons and introducing distance-dependent inter-neuron delays. These delays were shown to significantly improve ESN task performance. However, thus far, it is still unclear why distance-based delay networks (DDNs) perform better than ESNs. In this paper, we show that by optimizing inter-node delays, the memory capacity of the network matches the memory requirements of the task. As such, networks concentrate their memory capabilities to the points in the past which contain the most information for the task at hand. Moreover, we show that DDNs have a greater total linear memory capacity, with the same amount of non-linear processing power.

Splenectomy Does Not Increase Early Morbidity in Patients Undergoing Total Gastrectomy: A Case-Control Study with Propensity Score Analysis.

AIM: In gastric cancer (GC), D2 lymph node dissection is, alongside negative-margins gastrectomy, of paramount importance. There is a debate between Western and Eastern scientific communities concerning the risk-benefit balance with respect to splenectomy, as Western countries are inclined to perform spleen-preserving gastrectomy due to an increased risk for postoperative complications. In Eastern countries (such as Japan) this is not the case. Our study aimed to determine whether or not spleen-sacrificing total gastrectomy for GC was associated with a higher rate of early postoperative morbidity or mortality. METHOD: We performed a retrospective case-control study in which we included patients who underwent total gastrectomy with D2 lymphadenectomy for GC (stages I-III) with curative intent, in a single high-volume tertiary oncologic centre. We divided the cases into two groups: spleenpreserving (SP) and spleen-sacrificing (SS) and evaluated the early complications rate following surgery. Afterwards, we performed propensity score matching (PSM) and analysis of the two groups. Results: We included 74 patients, 29 in the SS group and 45 in the SP group. Fifteen cases (20.2%) developed early postoperative complications and the complication rate was 53% (n=8) in the SS group and 46% (n=7) in the SP group. The overall 30-day mortality rate was 2.7%. Conclusions: Splenectomy is not associated with increased early morbidity following total gastrectomy with D2 lymphadenectomy if performed by an experienced surgeon.

Evaluation of Changes in Circulating Cell-Free DNA as an Early Predictor of Response to Chemoradiation in Rectal Cancer-A Pilot Study.

Background and Objectives: The objective of this study was to investigate quantitative changes in cell-free DNA (cfDNA) found in the bloodstream of patients with locally advanced rectal cancer who received neoadjuvant long-course chemoradiation, assuming a change in DNA fragments release during therapeutic stress. Materials and Methods: This was a prospective observational study that involved 49 patients who had three distinct pathologies requiring neoadjuvant chemoradiation: 18 patients with breast cancer, 18 patients with cervical cancer, and 13 patients with rectal cancer. Both breast cancer and cervical cancer patients were used as a control groups. Breast cancer patients were used as a control group as irradiation targeted healthy tissue after the tumor resection (R0), while cervical cancer patients were used as a control group to evaluate the effect of chemoradiation regarding cfDNA in a different setting (squamous cell carcinomas) and a different tumor burden. Rectal cancer patients were the study group, and were prospectively evaluated for a correlation between fragmentation of cfDNA and late response to chemoradiation. Blood samples were collected before the initiation of treatment and after the fifth radiation dose delivery. cfDNA was quantified in peripheral blood and compared with the patients' clinicopathological characteristics and tumor volume. Conclusion: Thirteen patients with locally advanced rectal cancer (T3/T4/N+/M0) were included in the study, and all of them had their samples analyzed. Eight were male (61.54%) and five were female (38.46%), with an average age of 70.85 years. Most of the patients had cT3 (53.85%) or cT4 (46.15%) tumors, and 92.31% had positive lymph nodes (N2-3). Of the thirteen patients, only six underwent surgery, and one of them achieved a pathological complete response (pCR). The mean size of the tumor was 122.60 mm3 [35.33-662.60 mm3]. No significant correlation was found between cfDNA, tumor volume, and tumor regression grade. cfDNA does not seem to predict response to neoadjuvant chemoradiotherapy and it is not correlated to tumor volume or tumor regression grade.

Predictive and Prognostic Role of Neutrophil to Lymphocyte Ratio in Rectal Cancer: A Case Control Study with Propensity Score Analysis.

Background: Neutrophil to lymphocyte ratio (NLR) is promoted as a marker reflecting the antitumoral inflammatory response. Herein, we aim to assess whether NLR at the time of diagnosis can predict response to neoadjuvant therapy and long-term survival in a matched cohort of rectal cancer patients. Methods: This is a case control study on rectal cancer patients who underwent standard oncological treatment and had NLR sampled at each stage. ROC curve was used to establish the cut off value of NLR at diagnosis. Two groups (high and low NLR) were compared. Kaplan Meier overall and disease-free survival (DFS) analysis was done comparatively between two groups of patients: low and high NLR. Pearson and Log Rank tests were used to establish statistical significance. Propensity score matching (PSM) was performed, and all variables were compared again on the matched subgroups. Results: One hundred patients were included and 54 were compared again after PSM. NLR at diagnosis did not correlate with tumor regression grade (p=0.77). High NLR at diagnosis (NLR 2.58) was not found to be significantly associated with worse overall survival (p=0.096) or DFS (p=0.128). Similar results were achieved after PSM, except when stage III subgroups were compared, where higher NLR was associated with worse DFS (p=0.04), while results for OS were borderline (p=0.05). Conclusions: Overall, a pretherapeutic high NLR ( 2.58) was not found to predict survival or response do neoadjuvant therapy in patients with rectal cancer. However, a higher NLR may be associated with worse outcomes in advanced colorectal cancer.

The mesocolic apical fragment in complete mesocolic excision colectomies: Should it be analysed separately? A proof-of-concept study.

AIM: The aim of this work is to describe a protocol and assess the feasibility of harvesting and analysing the mesocolic apical fragment (MAF) for the presence of central lymph node (LN) metastasis and extra lymphatic free tumour cells in a random subgroup extracted from a cohort of complete mesocolic excision colectomies with central vascular ligation. METHOD: Forty-seven patients diagnosed with colorectal cancer were included. A 2/2 cm pyramid of tissue was cut around the central tie and sent for pathological examination. The MAF was sectioned into 16 slices. High-definition images were taken from the slices which were merged into a panoramic three-dimensional image of the MAF. The distribution of LNs in the MAF was quantified. Immunohistochemistry staining for cytokeratin 14 was used to identify isolated tumour cells and micrometastases in the extranodal tissue. RESULTS: No tumoural cells migrating through the apical zone, outside of the LNs, were identified. Margins of resection, mesocolic tissue and LNs were all negative in the subgroup of ultrastaged MAFs. The number of examined central LNs varied between 0 and 24, with positive MAF LNs being identified only in pN2 stages. The rate of positive apical LNs in our cohort was 4.2% (n = 2). CONCLUSIONS: The MAF can be easily extracted from standard specimens, allowing for accurate analysis of lymphatic and extra-nodal tumour cells on the central resection margins, in central LNs and in the apical mesocolic tissue. Future research on larger cohorts is required to establish if analysing the MAF has an impact on patient staging, prognosis and management.

Impact of Quilting Sutures on Surgical Outcomes After Mastectomy: A Systematic Review and Meta-Analysis.

BACKGROUND: Seroma after mastectomy and/or axillary lymph node dissection (ALND) is among the most common issue surgeons have to face in the early postoperative management of breast cancer. Using quilting sutures (QS) to aid in tissue approximation and decrease dead space is proposed as a simple technique to reduce seroma rate. We aimed to perform a systematic review, and analyse, in a meta-analytical model, the role of QS in improving wound outcomes and decrease volume, duration of drainage, and length of stay in hospital. METHODS: The study was registered with PROSPERO. A systematic search of the PubMed, EMBASE, and SCOPUS databases was performed for all comparative studies examining surgical outcomes in patients who underwent QS versus conventional closure (CC) after mastectomy ± ALND. RESULTS: Twenty-one studies with a total of 3473 patients (1736 in the study group and 1737 in the control group) were included based on the selection criteria. The study group showed significantly lower rates of seroma (p < 0.00001), total volume of drainage (p < 0.0001), days to drain removal (p < 0.00001), and length of stay (p < 0.00001) compared with the control group, while wound complication rates (surgical site infection, flap necrosis, hematoma, skin dimpling) were comparable between the two groups. CONCLUSIONS: QS are a reliable intraoperative technique that decrease seroma formation, volume of postoperative drainage, duration of drainage and length of hospital stay, and should be considered in mastectomies with or without ALND.