Machine learning for intraoperative prediction of viability in ischemic small intestine.

OBJECTIVE: Evaluation of intestinal viability is essential in surgical decision-making in patients with acute intestinal ischemia. There has been no substantial change in the mortality rate (30%-93%) of patients with acute mesenteric ischemia (AMI) since the 1980s. As the accuracy from the first laparotomy alone is 50%, the gold standard is a second-look laparotomy, increasing the accuracy to 87%-89%. This study investigates the use of machine learning to classify intestinal viability and histological grading in pig jejunum, based on multivariate time-series of bioimpedance sensor data. APPROACH: We have previously used a bioimpedance sensor system to acquire electrical parameters from perfused, ischemic and reperfused pig jejunum (7 + 15 pigs) over 1-16 h of ischemia and 1-8 h of reperfusion following selected durations of ischemia. In this study we compare the accuracy of using end-point bioimpedance measurements with a feedforward neural network (FNN), versus the accuracy when using a recurrent neural network with long short-term memory units (LSTM-RNN) with bioimpedance data history over different periods of time. MAIN RESULTS: Accuracies in the range of what has been reported clinically can be achieved using FNN's on a single bioimpedance measurement, and higher accuracies can be achieved when employing LSTM-RNN on a sequence of data history. SIGNIFICANCE: Intraoperative bioimpedance measurements on intestine of suspect viability combined with machine learning can increase the accuracy of intraoperative assessment of intestinal viability. Increased accuracy in intraoperative assessment of intestinal viability has the potential to reduce the high mortality and morbidity rate of the patients.

Ischemia/reperfusion injury in porcine intestine - Viability assessment.

AIM: To investigate viability assessment of segmental small bowel ischemia/reperfusion in a porcine model. METHODS: In 15 pigs, five or six 30-cm segments of jejunum were simultaneously made ischemic by clamping the mesenteric arteries and veins for 1 to 16 h. Reperfusion was initiated after different intervals of ischemia (1-8 h) and subsequently monitored for 5-15 h. The intestinal segments were regularly photographed and assessed visually and by palpation. Intraluminal lactate and glycerol concentrations were measured by microdialysis, and samples were collected for light microscopy and transmission electron microscopy. The histological changes were described and graded. RESULTS: Using light microscopy, the jejunum was considered as viable until 6 h of ischemia, while with transmission electron microscopy the ischemic muscularis propria was considered viable until 5 h of ischemia. However, following ≥ 1 h of reperfusion, only segments that had been ischemic for ≤ 3 h appeared viable, suggesting a possible upper limit for viability in the porcine mesenteric occlusion model. Although intraluminal microdialysis allowed us to closely monitor the onset and duration of ischemia and the onset of reperfusion, we were unable to find sufficient level of association between tissue viability and metabolic markers to conclude that microdialysis is clinically relevant for viability assessment. Evaluation of color and motility appears to be poor indicators of intestinal viability. CONCLUSION: Three hours of total ischemia of the small bowel followed by reperfusion appears to be the upper limit for viability in this porcine mesenteric ischemia model.

Small intestinal ischemia and reperfusion-bioimpedance measurements.

OBJECTIVE: Trans-intestinal bioimpedance measurements have previously been used to investigate changes in electrical parameters during 6 h of ischemia in the small intestine. Knowledge is lacking regarding the time course of trans-intestinal bioimpedance parameters during reperfusion. As reperfusion is an important part in the clinical treatment of intestinal ischemia, we need to know how it affects the bioimpedance measurements. APPROACH: We performed bioimpedance measurements, using a two-electrode setup on selected segments of the jejunum in 15 pigs. A controlled voltage signal was applied while measuring the resulting current. In each pig, five or six 30 cm segments of the jejunum were made ischemic by clamping the mesenteric arteries and veins creating segments with ischemia from 1-16 h duration. Reperfusion was initiated at selected time intervals of ischemia, and measured for 5-15 h afterwards. MAIN RESULTS: The tan δ parameter (loss tangent) was different (p < 0.016) comparing ischemic and control tissue for the duration of the experiment (16 h). Comparing the control tissue 30 cm from the ischemic area with the control tissue 60 cm from the ischemic tissue, we found that the mean tan δ amplitude in the frequency range (3900-6300 Hz) was significantly higher (p < 0.036) in the proximal control after 10 h of experiment duration. After reperfusion, the time development of tanδm (loss tangent maximum over a frequency range) amplitude and frequency overlapped and periodically increased above the tanδm in the ischemic intestine. Dependent on the ischemic duration pre-reperfusion, the initial increase in tan δ stabilizes or increases drastically over time, compared to the tan δ amplitude of the ischemic tissue. SIGNIFICANCE: As during ischemia, the electrical parameters during reperfusion also follow a characteristic time-course, depending on the ischemic exposure before pre-reperfusion. The temporal changes in electrical parameters during small intestinal ischemia followed by reperfusion provides important information for assessment of tissue injury.

Ischemic small intestine-in vivo versus ex vivo bioimpedance measurements.

OBJECTIVE: Bioimpedance has been used to investigate changes in electrical parameters during ischemia in various tissues. The small intestine is a multi-layered structure, with several distinct tissue types, and ischemia related changes occur at different times in the different intestinal layers. When investigating how the electrical properties in the small intestine is affected by ischemia, some researchers have used ex vivo models while others have used in vivo models. In this study, we compare ischemic time development of electrical parameters in ischemic in vivo versus ex vivo small intestine. APPROACH: Measurements were performed using a two-electrode setup, with a Solartron 1260/1294 impedance gain-phase analyser. Electrodes were placed on the surface of ischemic pig jejunum, applying a voltage and measuring the resulting electrical admittance. In each pig, 4 segments of the jejunum were made ischemic by clamping the mesenteric arteries and veins, resulting in a 30 cm central zone of warm ischemia and edema. The in vivo part of the experiment lasted 10 h, after which 3 pieces of perfused small intestine were resected, stored in Ringer-acetat at 38 °C, and measured during a 10 h ex vivo experiment. Main results and significance: We found significant differences (p < 0.0001) between the values of electric parameters when comparing the in vivo and ex vivo measurements as a function of ischemic time development. We also observed some similarities in the trends. In vivo, we measured an overall decrease in impedance during the duration of the experiment, probably as a result from the formation of edema. Ex vivo, the low frequency impedance increased initially for approximately 3 h before starting to decrease.

Dissecting the effects of lipopolysaccharides from nonlipopolysaccharide molecules in experimental porcine meningococcal sepsis.

OBJECTIVE: To dissect the in vivo responses to lipopolysaccharide compared with nonlipopolysaccharide structures of whole meningococci. DESIGN: Comparative experimental study. SETTING: University hospital with an animal intensive care unit and laboratory. SUBJECTS: Twenty-four anesthetized healthy Norwegian landrace pigs of 30 kg (+/- 2.5 kg) grouped into two test groups and one control group. INTERVENTIONS: Exponentially increasing numbers of Neisseria meningitidis H44/76 (NmLPS+) or a knockout mutant of H44/76 completely lacking lipopolysaccharide (NmLPS-) were infused intravenously to the pigs. MEASUREMENTS AND MAIN RESULTS: Physiological and hematologic parameters were continuously recorded and biochemical analyses were performed in batch after completion. Systemic vascular resistance, cardiac index and lactate changed significantly more in the NmLPS+ than in the NmLPS- group (p < .05). Mean pulmonary artery pressure increased early in the NmLPS+ and late in the NmLPS- group, but finally reached equally high values. Capillary leakage (fluid requirement, plasma albumin loss, organ wet/dry ratio) was more prominent in the NmLPS+ group (p < .05). Leukocytes were depleted in a highly lipopolysaccharide-dependent manner (p < .001). Thrombin-antithrombin complexes and plasminogen activator inhibitor-1 increased 2.5 to five times more in the NmLPS+ group (p < .05). Maximum cytokine concentrations in plasma were markedly higher in the NmLPS+ group (p < .05): tumor necrosis factor-alpha (40 times), interleukin-1beta (40 times), interleukin-6 (13 times), and interleukin-10 (four times). Interleukin-12 increased only in the NmLPS+ group. CONCLUSION: This large animal model, which simulates human disease well, confirms the potency of lipopolysaccharide but provides clear evidence that nonlipopolysaccharide molecules induce cardiovascular and hematologic changes quite similar to those caused by lipopolysaccharide. In general, 10- to 20-fold higher doses of the lipopolysaccharide-deficient mutant were required to induce the same degree of pathophysiological changes. Endotoxic activity of Gram-negative bacteria should no longer be attributed solely to the activity of lipopolysaccharide.

Patient-reported outcome after randomization to on-pump versus off-pump coronary artery surgery.

BACKGROUND: Clinical experience with off-pump coronary artery bypass surgery raises the question of a patient experienced benefit compared with on-pump surgery. This prospective and randomized study compared patient-reported outcome between surgical groups, as change scores at 3 months after surgery and longitudinally as time-averaged change from baseline through the first year after surgery. METHODS: In all, 120 patients were randomly assigned to on- or off-pump coronary artery surgery. A questionnaire for patient self-report of angina (Canadian Cardiovascular Society scale), health status (Short Form 36, sleep and sexual difficulty), and overall quality of life (Quality of Life Scale) was administered at baseline and at 3, 6, and 12 months after surgery. RESULTS: Patient groups were comparable with regard to age, symptoms, comorbidity, and surgical characteristics. Both groups experienced a median of two classes relief of angina at 3 months (p < 0.0005), maintained throughout follow-up. Paired t tests revealed significant improvement on all Short Form 36 subscales at 3 months after surgery, with the exception of physical role functioning in the on-pump group. No independent main effects of surgical group were observed in the between-groups covariance models. The longitudinal effect of sex was significant in four Short Form 36 subscales: physical functioning, bodily pain, and role limitation due to physical or emotional problems. Overall quality of life scores were stable in both groups. CONCLUSIONS: Both on-pump and off-pump patients reported less angina and improved health status after surgery. There were no significant differences between surgical groups in health status or overall quality of life, neither cross-sectionally nor longitudinally.