Predicting Intestinal Viability After Acute Mesenteric Ischemia by Photoacoustic Imaging: An Experimental Study in Rats


 Our purpose was to assess the ability of photoacoustic imaging (PAI) to objectively predict ischemic intestinal pathological damage during surgery for acute mesenteric ischemia (AMI). PAI uses laser light to detect relative amounts of oxygenated and deoxygenated hemoglobin in intestinal tissue. In 18 rats, AMI was induced by clamping the mesenteric and marginal vessels of the ileum for 0 minutes in the sham group (n = 6), 30 minutes in the mild group (n = 6), and 180 minutes in the severe group (n = 6). After 60 minutes of reperfusion, the damage of the intestine was evaluated pathologically. Oxygenation of the intestine was monitored throughout the procedure in real time by the PAI system and compared between the mild and severe groups. All rats showed irreversible (i.e. transmucosal or transmural infarction) damage in the severe group. After reperfusion, the oxygenation in the mild group recovered immediately and was significantly higher than in the severe group at 1, 5, 10, 30, and 60 minutes (P = .007, 014, 016, 008, and 003, respectively). Oxygenation showed a significant strong negative correlation with the pathological severity. In conclusion, PAI could objectively predict irreversible ischemic damage immediately after reperfusion, which potentially prevents inadequate surgery.


Introduction
During surgery for acute mesenteric ischemia (AMI), since there is no reliable objective marker of intestinal viability to determine the resection margin, it is assessed subjectively by the surgeon. However, this requires a high level of clinical experience 1-4 . In recent years, photoacoustic imaging (PAI), also known as optoacoustic imaging, has been demonstrating its usefulness in a variety of elds, including clinical practice [5][6][7][8][9][10][11][12][13] . It is an emerging realtime in vivo imaging modality that combines optical imaging contrast with ultrasonic spatial resolution. PAI uses the near-infrared spectrum, which has deeper penetration and visualizes deeper tissue in vivo in real time. Light energy is selectively absorbed by endogenous and exogenous chromophores in tissue causing rapid thermoelastic expansion, which generates broadband acoustic waves that can be detected by ultrasound transducers 14 . Differing absorption spectra of oxygenated hemoglobin (HbO 2 ) and deoxyhemoglobin (Hb) with dual-wavelength PAI allow quanti cation of the blood oxygen level. The usefulness of oxygenation mapping in organ ischemia has also been reported 15,16 .
With PAI, oxygenation of ischemic intestine after reperfusion can be evaluated noninvasively during surgery. The purpose of this study was to evaluate the usefulness of oxygenation mapping by PAI to objectively and quantitatively predict ischemic intestinal pathological damage during surgery for acute mesenteric ischemia (AMI).

Page 3/16
Laboratory tests Results of laboratory tests are shown in Fig. 1(A). Local capillary lactate was signi cantly higher in the severe group (14.2 ± 8.6 mmol/L) than in the mild group and the sham group (4.1 ± 3.1 and 1.0 ± 0.1 mmol/L, P = .01 and .002, respectively). There was no signi cant difference between the mild group and the sham group (P = .58). Systemic lactate was signi cantly higher in the severe group and in the mild group (0.6 ± 0.1 and 0.7 ± 0.2 mmol/L) than in the sham group (0.4 ± 0.1 mmol/L, P = .03, 001, respectively), but there was no signi cant difference between the severe group and the mild group (P = .27). LDH (severe, 431.3 ± 131.0; mild, 188.7 ± 34.9; sham, 139.7 ± 18.5 U/L; respectively) and CK (severe, 406.2 ± 144.5; mild, 143.3 ± 11.8; sham, 125.8 ± 23.1 U/L; respectively) were signi cantly higher in the severe group than in the others (all P < .001, for details, see Fig. 1(A)), but there was no signi cant difference between the mild and the sham groups (P = .54 and .93). AST did not show a signi cant difference between any two groups (severe, 73.5 ± 20.5; mild, 64.3 ± 8.9; sham, 63.0 ± 5.9 U/L; respectively. P = .36).

Time course of oxygenation in AMI
Before clamping, average oxygenation was greater than 70%. All rats showed a rapid decrease of oxygenation after clamping, and then reached a plateau (< 30%) within ve minutes. The time course of the oxygenation after clamping is shown in Fig. 1(B).
The appearance of ischemic intestine was dark, and it tended to be darker in the severe group, but it was not always distinguishable between the two ischemic groups (Fig. 2).  Table 1 show the time course of oxygenation in two groups after reperfusion.
Before removing the vascular clamps, there was no signi cant difference between the two groups. Once the vascular clamps were removed, oxygenation recovered to the baseline level (about 65-75%) immediately in the mild group. On the other hand, oxygenation recovered slightly but remained lower in the severe group. At 1, 5, 10, 30, and 60 minutes after the removal of vascular clamps, oxygenation was signi cantly lower in the severe group than in the mild group. Note -After reperfusion, the time course of oxygenation was clearly different between the mild group and the severe group. Oxygenation was signi cantly higher in the mild group than in the severe group at 1, 5, 10, 30, and 60 min. All values are shown as means (standard error).

Pathological ndings
In the severe group, all tissue samples showed irreversible damage, and the average score was 7.2 ± 0.4, signi cantly higher than in the other groups (both P < .001). In the mild group, the average score was 1.2 ± 1.8, and in the sham group, the average score was 0.
Correlation analysis between oxygenation and the pathological score Correlation analysis between oxygenation and the pathological score was performed ( Fig. 3(C), Table 2). Table 2 shows Spearman's correlation coe cient (r s -value) along with the corresponding P-values for the correlation between oxygenation and the pathological score, showing a signi cant strong negative correlation at each time point after reperfusion. Correlation analysis between local capillary lactate and the pathological score was also performed, but it did not show a signi cant correlation (P = 0.06). Note -The P value and r s value between oxygenation and the pathological score are shown. There was a signi cant, strong negative correlation between oxygenation and the pathological score at 1, 5, 10, 30, and 60 min.

Discussion
During AMI surgery, intestinal viability is subjectively assessed by surgeons and is highly dependent on their experience. Karliczek et al. evaluated surgeons' predictive accuracy for anastomotic leakage in gastrointestinal surgery, but the predictive value was low in a prospective, clinical study. Incorrect evaluation may result in bowel stricture, obstruction, necrosis, anastomotic leakage, and short bowel syndrome 2,17 . Thus, the usefulness of oxygenation mapping by PAI as an objective, quantitative, in vivo method to assess intestinal viability in AMI was evaluated. In the present study, there was a signi cant difference in the time course of oxygenation immediately after reperfusion between the two ischemic groups with different prognoses. In addition, oxygenation in the intestine at each time point showed a signi cant strong negative correlation between the pathological scores, even at 1 minute after reperfusion. To the best of our knowledge, this is the rst report to evaluate intestinal viability in AMI with oxygenation mapping by PAI.
Oximetry with Photoacoustic imaging Singh et al. showed that the serosal oxygenation measurements were not sensitive to ischemia 18 . It has been estimated that blood ow in the mucosal and submucosal layers accounts for 70% of total blood ow in intestinal tissues; therefore, microcirculation should be evaluated in the mucosal and submucosal layers 19 . There are various methods to evaluate intestinal viability during surgery, and most of them re ect oxygenation and perfusion from the serosal surface, but none of them has become a clinical standard due to various de ciencies 2,20 . In recent years, PAI has been demonstrating its clinical usefulness, which can image deeper tissue and can be performed at the bedside in real time 5,8,10,11 . It is worth mentioning that PAI can be performed during surgery as well as ultrasound. With this method, oxygenation of the layers including the mucosal layer can be evaluated noninvasively. In the present study with the preclinical PAI system, layers including the mucosal and submucosal layers were evaluated.

Sequential real-time evaluation of oxygenation
In the present study, before inducing AMI, average oxygenation of normal ileum was between 70% and 80%, which was compatible with the previous report using visual light spectroscopy 17 . In the early stage of AMI, oxygenation reached a plateau within ve minutes. From the present result, intestinal viability cannot be predicted by the absolute oxygenation value itself before reperfusion. This nding is clinically acceptable. In clinical practice, decreased staining of the intestinal wall on contrast-enhanced computed tomography is a key nding for diagnosing AMI, but an intestinal segment with decreased staining is not always irreversibly damaged after reperfusion.
The decrease in oxygenation can be caused either by a decrease in arterial blood with rich oxygenation or by possible pooling of venous blood. The signi cantly low oxygenation after reperfusion in the severe AMI group may due to vasospasm, thrombosis, and congestion in damaged intestinal tissue 21 .
The present study has limitations. First, in strangulated bowel obstruction in clinical practice, venous ow is compromised prior to arterial ow. In the present model, both arterial and venous ow were obstructed simultaneously. Second, the reason why oxygenation decreased was not speci cally clari ed. In addition, the regions of interest in intestinal tissue on PAI could not be maintained continuously in a particular (mucosal and submucosal) layer under respiratory motion, due to the small size of rats; thus, oxygenation was evaluated using regions of interest including all layers.

Conclusion
Real-time PAI could objectively and quantitatively predict irreversible ischemic tissue damage immediately after reperfusion in AMI, and it can be performed at the bedside and potentially prevent inadequate surgery.

Animal models
Eighteen male Sprague-Dawley rats provided by Charles River Laboratories Japan (Yokohama, Japan) were studied. All animal experiments were performed in accordance with the guidelines of our institution and approved by our animal research committee (study protocol ID: AP194060). Rats were maintained in a speci c pathogen-free facility. All rats weighed approximately 600 g and had ad libitum access to food and water. AMI was established by the procedure described below.

Rat model of AMI
After 24 hours of fasting, the abdomen was shaved. A 3-cm midline incision was made, and a 15-cm length of ileum from the ileocecal valve was exposed. Blood samples were obtained from the femoral vein to evaluate systemic lactate, lactate dehydrogenase (LDH), aspartate transaminase (AST), and creatine phosphokinase (CK) levels. Subsequently, local capillary lactate was evaluated by puncturing the intestine. The rats were randomly assigned into three groups in advance. The rst group was the sham group (n = 6), which was designed to assess the effect of anesthesia and the invasiveness of the operation to standardize the study in terms of ischemia. The second group was the mild AMI group (n = 6), in which mesenteric arteries and veins and marginal arteries and veins were clamped with vascular clamps, and mesenteric ischemia was then induced in 15 cm of ileum from the ileocecal valve (strangulated bowel obstruction model; Fig. 4); after 30 minutes of ischemia, the clamps were removed, and blood samples were obtained, and local capillary lactate was again evaluated; after 60 minutes of reperfusion, a tissue sample was obtained for pathological evaluation. The third group was the severe AMI group (n = 6), which was the same as the mild AMI group, but with 180 minutes of ischemia. During AMI and reperfusion, oxygenation of ischemic intestine was monitored by a commercially available PAI system Vevo LAZR /Vevo 2100 (FUJIFILM VisualSonics Inc., Toronto, Canada, Fig. 5(A)).
Induction of anesthesia was performed with 5% iso urane, and it was then maintained by intraperitoneal injection of a mixture of medetomidine hydrochloride (0.15 mg/kg), midazolam (2.0 mg/kg), and butorphanol tartrate (2.5 mg/kg). Vital signs (heart rate, respiratory rate, peripheral oxygen saturation) were continuously monitored by MouseOx PLUS (Starr Life Sciences Co., PA, USA) to exclude hypoventilation. The neck of each rat was also shaved, and a clip sensor was a xed. During the procedure, the rats were kept on a heating pad to avoid hypothermia.

Ultrasound and PAI
PAI was performed using a Vevo LAZR/2100 imaging system with an LZ-550 linear-array transducer (256 elements, center frequency 40 MHz, bandwidth 27 MHz) used to acquire all images. The tunable laser supplied 10-20 mJ per pulse over the 680-970 nm wavelength range, with a pulse repetition frequency of 10 Hz. Once initialized, the system was switched to "Oxy-Hemo" mode to obtain parametric maps of the oxygenation using the following parameters: depth 6 mm; width 14.08 mm; and wavelength 750 and 850 nm. The oxygenation mapping was shown in parallel with B-mode images (Fig. 5(B)). Before the removal of vascular clamps, and 1, 5, 10, 30 and 60 minutes after removal, ve regions of interest were placed on layers including the mucosal layer, submucosal layer, muscular layer, and subserosal layer, and oxygenation was measured for 10 seconds and averaged. Oxygenation of the mild and severe groups was compared.

Pathological ndings
At the end of each experiment, a tissue sample was obtained for histopathological study. Tissues were stained with hematoxylin and eosin and evaluated by a pathologist blinded to group assignment. The sections from the intestine were classi ed according to the degree of tissue injury in accordance with Chiu's score classi cation including modi cations proposed by Park 22,23 . This Park/Chiu's classi cation of small intestinal injury consists of values from 0 to 8, which are shown in Table 3. Samples of pathological ndings are shown in Fig. 6. Scores of 7 and 8 were considered to indicate irreversible ischemic damage.  Figure 1 (A) The results of laboratory tests in the three groups (n = 6 rats per group) (B) Time course of mean oxygenation in the early stage of AMI A rapid decrease of oxygenation from baseline (> 70%) after clamping is clearly visualized. The oxygenation reaches a plateau within 5 minutes. Error bars represent standard errors of the mean (n = 12 rats). Note -Sham, sham group; Mild, mild acute mesenteric ischemia group; Severe, severe acute mes-enteric ischemia group; LDH, lactate dehydrogenase; AST, aspartate transaminase; CK, creatine phosphoki-nase; ns, not signi cant; *, p < .05; **, p < .005; ***, p < .0005; ****, p < .0001; AMI, acute mesenteric is-chemia Figure 2 Gross appearance of ischemic ileum in the mild and severe groups In some cases, the severity of AMI is indistinguishable by the gross appearance alone. Note -AMI, acute mesenteric ischemia The correlation between oxygenation and the pathological score with the best-t line with 95% con -dence band After reperfusion, there is a negative linear relationship between oxygenation and the pathological score. The correlation coe cient (rs) and p values are shown. Note -AMI, acute mesenteric ischemia; Mild, the mild AMI group; Severe, the severe AMI group Figure 4 Strangulated bowel obstruction model in rats Acute mesenteric ischemia was induced in 15 cm of ileum from the ileocecal valve using vascular clamps. Mesenteric arteries and veins and marginal arteries and veins were clamped.