New perspectives in the prediction of postoperative complications for high-risk ulcerative colitis patients: machine learning preliminary approach

OBJECTIVE

Patients with acute severe and medical refractory ulcerative colitis have a high risk of postoperative complications after total abdominal colectomy (TAC). The objective of this retrospective study is to use machine learning to analyze and predict short-term outcomes.

PATIENTS AND METHODS

32 patients with ulcerative colitis were treated with total abdominal colectomy between 2011 and 2017. Biographical data, preoperative therapy, blood chemistry, nutritional status, surgical technique, blood transfusion and preoperative length of stay were the features selected for the statistical analyses and were used as input for the machine learning algorithms to predict the rate of complications.

RESULTS

Traditional statistical analysis showed an overall postoperative morbidity rate of 34% and a mortality rate of 3%. Preoperative low serum albumin levels (<2.5 g/dL) were related to a higher risk of minor infectious complications with statistical significance (p<0.05). Preoperative length of stay (>4 days), blood transfusions (≥1 unit) and body temperature (≥37.5°C) demonstrated a major impact on infectious morbidity with statistical significance (p<0.05). Patients treated with steroids and rescue therapy presented a higher risk of minor infectious complications (p<0.05). Evaluating only preoperative features, machine learning algorithms were able to predict minor postoperative complications with a high strike rate (84.3%), high sensitivity (87.5%) and high specificity (83.3%) during the testing phase.

CONCLUSIONS

Machine learning is demonstrated to be useful in predicting the rate of minor postoperative complications in high-risk ulcerative colitis patients, despite the small sample size. It represents a major step forward in data analysis by implementing a retrospective study from a prospective point of view.

Role of non-genomic androgen signalling in suppressing proliferation of fibroblasts and fibrosarcoma cells

The functions of androgen receptor (AR) in stromal cells are still debated in spite of the demonstrated importance of these cells in organ development and diseases. Here, we show that physiological androgen concentration (10 nM R1881 or DHT) fails to induce DNA synthesis, while it consistently stimulates cell migration in mesenchymal and transformed mesenchymal cells. Ten nanomolar R1881 triggers p27 Ser10 phosphorylation and its stabilization in NIH3T3 fibroblasts. Activation of Rac and its downstream effector DYRK 1B is responsible for p27 Ser10 phosphorylation and cell quiescence. Ten nanomolar androgen also inhibits transformation induced by oncogenic Ras in NIH3T3 fibroblasts. Overexpression of an AR mutant unable to interact with filamin A, use of a small peptide displacing AR/filamin A interaction, and filamin A knockdown indicate that the androgen-triggered AR/filamin A complex regulates the pathway leading to p27 Ser10 phosphorylation and cell cycle arrest. As the AR/filamin A complex is also responsible for migration stimulated by 10 nM androgen, our report shows that the androgen-triggered AR/filamin A complex controls, through Rac 1, the decision of cells to halt cell cycle and migration. This study reveals a new and unexpected role of androgen/AR signalling in coordinating stromal cell functions.

Inhibition of the SH3 domain-mediated binding of Src to the androgen receptor and its effect on tumor growth

In human mammary and prostate cancer cells, steroid hormones or epidermal growth factor (EGF) trigger association of the androgen receptor (AR)-estradiol receptor (ER) (alpha or beta) complex with Src. This interaction activates Src and affects the G1 to S cell cycle progression. In this report, we identify the sequence responsible for the AR/Src interaction and describe a 10 amino-acid peptide that inhibits this interaction. Treatment of the human prostate or mammary cancer cells (LNCaP or MCF-7, respectively) with nanomolar concentrations of this peptide inhibits the androgen- or estradiol-induced association between the AR or the ER and Src the Src/Erk pathway activation, cyclin D1 expression and DNA synthesis, without interfering in the receptor-dependent transcriptional activity. Similarly, the peptide prevents the S phase entry of LNCaP and MCF-7 cells treated with EGF as well as mouse embryo fibroblasts stimulated with androgen or EGF. Interestingly, the peptide does not inhibit the S phase entry and cytoskeletal changes induced by EGF or serum treatment of AR-negative prostate cancer cell lines. The peptide is the first example of a specific inhibitor of steroid receptor-dependent signal transducing activity. The importance of these results is highlighted by the finding that the peptide strongly inhibits the growth of LNCaP xenografts established in nude mice.

Evaluation of three oral dosages of ondansetron in the prevention of nausea and emesis associated with cyclophosphamide-doxorubicin chemotherapy

We assessed the antiemetic efficacy and safety of three different oral doses of ondansetron (GR 38032F), a novel serotonin type-3 receptor antagonist, in three consecutive series of 20 breast cancer patients receiving cyclophosphamide-doxorubicin-based chemotherapy for the first time. Patients received oral doses of 8 mg, 4 mg, or 1 mg of ondansetron three times daily for 2 days, with the first dose given 30 minutes before the cyclophosphamide infusion. We then evaluated the efficacy of a conventional antiemetic regimen of intravenous lorazepam, metoclopramide, and diphenhydramine given before chemotherapy and 10 mg prochlorperazine given orally twice on study day 1 and three times on study day 2 in a fourth series of 20 patients with comparable characteristics. The number of emetic episodes, assessment of nausea and appetite, and adverse events were recorded throughout the 2-day study period. Pretreatment and posttreatment clinical laboratory data were also collected. No emesis was observed during the 2-day study period in 17 (85%), 13 (65%), and 11 (55%) patients treated with 8-mg, 4-mg, and 1-mg ondansetron doses, respectively, and in seven (35%) patients who received conventional therapy. The incidence and intensity of nausea were lower with increasing doses of ondansetron and were lower than in the conventional group. Ondansetron-related side effects were generally mild and reversible and did not appear to increase in a dose-dependent manner. These effects included headache, stomach cramps, diarrhea, fatigue, and elevated serum transaminase concentrations. One patient who received three 1 mg doses of ondansetron experienced tremors and muscle twitching. Oral ondansetron is an effective and safe antiemetic for patients receiving noncisplatin cyclophosphamide-doxorubicin-based chemotherapy, and its antiemetic activity appears to be dose-related.

GR 38032F (ondansetron), a selective 5HT3 receptor antagonist, slows colonic transit in healthy man

The newly recognized class of 5-hydroxytryptamine receptors (5HT3) may be involved in the induction of nausea, since their pharmacological antagonists are effective against emesis induced by chemotherapy. 5HT3 receptors are present on enteric neurons, and 5HT3 blockers may produce mild constipation; we thus hypothesized that 5HT3 receptors would modulate colonic motility. To determine if GR 38032F, a selective 5HT3 antagonist known to have antiemetic effects, influences colonic transit in health, a randomized, double-blind, placebo-controlled crossover study was performed. Using a radiopaque marker technique, colonic transit was quantified in 39 healthy volunteers (19 men, 20 nonpregnant women) 18-70 years of age. On a standard 25-g fiber diet, 16 mg of GR 38032F was given orally thrice daily. Gastrointestinal peptides (peptide YY, human pancreatic polypeptide, neurotensin, motilin, gastrin-cholecystokinin, substance P) were also measured in plasma fasting and postprandially. Mean total colonic transit time on placebo was 27.8 hr, while on GR 38032F it was 39.1 hr (P less than 0.0005). Transit times through the left colon (P less than 0.0005) and rectosigmoid (P less than 0.05) were prolonged by the drug, but right colonic transit was not significantly altered. Transit times did not correlate with age or gender, but subjects with shorter transit times were significantly more affected than were those with longer transit times. The peak release of peptide YY was minimally decreased following GR 38032F (P less than 0.01), but the peak and integrated postprandial responses of human pancreatic polypeptide, neurotensin, motilin, gastrin-cholecystokinin, and substance P were not significantly altered by the drug.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of selective 5HT3 antagonist (GR 38032F) on small intestinal transit and release of gastrointestinal peptides

Antagonists of 5-hydroxytryptamine type 3 (5HT3) receptors reduce the nausea induced by cisplatinum, but the effects of these agents on 5HT3 receptors in the human gut remain to be defined. We examined the actions of one of these drugs (Glaxo GR 38032F) on small intestinal transit and mouth-to-cecum transit times in healthy man. We also quantified its effects on the release of peptide YY (PYY), neurotensin, human pancreatic polypeptide, gastrin-cholecystokinin, and motilin. Ten healthy volunteers were enrolled in a randomized, double-blind, placebo-controlled crossover study. Following a single intravenous dose of GR 38032F (0.15 mg/kg), we measured the time to appearance in plasma of sulfapyridine after injection of salicylazosulfapyridine into the duodenum. This was used as a measure of duodenocecal transit. The appearance of hydrogen in breath after ingestion of a meal containing lactulose was also correspondingly used to quantify the mouth-to-cecum transit of the "head" of the meal. Gastrointestinal hormones were assayed in plasma by specific RIAs; samples were drawn fasting (10 min after injection) and after breakfast (358 calories: 15.7 g protein, 55.4 g carbohydrate, 8.1 g fat). The postprandial integrated response and peak release of PYY was decreased by GR 38032F. There was also a trend for the peak release of neurotensin to be reduced. GR 38032F did not significantly alter small intestinal transit times or mouth-to-cecum transit times. We conclude that GR 38032F does not have a major effect on small intestinal transit in health.