Imaging Evolution Of Brain Ischemic Penumbra: Experimental Study

Background Recently, the incidence of brain ischemic vascular diseases has been increased progressively in our country, which has led to severe social burden for its high mortality and disability incidence. Multiple experiments and clinical series confirmed that the degree of ischemic brain tissue within infarct lesion was variable, i.e., ischemic penumbra (IP) was present. IP is thought of as the target of the treatment of acute infarction, so much more studies have been focus on it. The purpose in our study is to determine the value of different imaging methods in IP.Part I: The applications of MR Diffusion Imaging in rat brain ischemic penumbraObjective Wistar rats were used to make middle cerebral artery occlusion (MCAO) brain ischemia-reperfusion model. Our purpose is to study the evolution of ischemic lesion on DWI and DTI and the values of DWI and DTI in IP.Materials and Methods Two hundred mature male Wistar rats were randomly divided into five groups, permanent ischemia group and four reperfusion groups after 0.5h, 1.5h, 2.5h and 3.5h respectively. MCAO model was made by modified Zea-Longa method. MR imaging was performed with GE 1.5T Twin Speed Infinity with Excite I magnetic resonance system. Three inch surface coil was used. The sequences included axial DWI, DTI and T2WI, which were performed 30min, 1h, 2h, 3h, 4h, 5h, 6h, 9h, 12h, 15h and 24h after MCAO respectively. Four region of interest (ROI) were measured. Point 1, 2 and 3 were located within the high signalintensity region on DWI. Point 1 was infarct core, Point 3 was infarct margin, and Point 2 was the midpoint between Point 1 and 3. Point 4 demonstrated normal signal intensity just adjacent to high signal intensity on DWI. Parameters measured included apparent diffusion coefficient (ADC), average diffusion coefficient (DCavg), fractional anisotropy (FA), relative anisotropy (RA) and volume ratio (VR). The relative ratios between infarct lesions and collateral normal brain were calculated. Statistical analysis was performed by SPSS 11.0 software package. Differences were considered statistically significant at P less than 0.05. The rats were decapitated after MPI examination, and HE, TTC and COX-2 pathologic and immunohistochemical stain were performed.Results (1) The diffusion parameters of ischemic lesions in rat MCAO model evoluted regularly within 24 hours after occlusion. ADC and DCavg demonstrated double-peak changes: deceased first, elevated slightly from 6 hours to 9 hours, and declined irreversibly. FA, 1-VR and RA demonstrated higher than normal before 9 hours and then lower than normal. (2) There were significant differences of ADC and DCavg between Point 1 and Point 2 within 3 hours, between Point 1 and Point 3 within 9 hours, between Point 2 and Point 3 from 3h to 9h (PO.05). There were significant differences of FA, 1-VR and RA between Point 1 and Point 2 at 5h and 6h, between Point 1 and Point 3 at 5h and 6h, between Point 2 and Point 3 within 6 hours (PO.05). All the parameters of Point 4 statistically differed from those of Point 1, 2 and 3 within 24 hours (P<0.05). (3) None of the parameters were significantly different between reperfusion groups after 2.5h and 3.5h and permanent ischemia group (P>0.05). The parameters of reperfusion groups after 0.5h and 1.5h significantly differed from those of reperfusion groups after 2.5h and 3.5h and permanent ischemia group (PO.05). (4) The infarct area between TTC stain and high signal intensity on DWI was significant difference within 6 hours (PO.05) but notafter 6h (P>0.05). The infarct areas in reperfusion groups after 0.5h at 24h significantly differed from those of reperfusion groups after 2.5h and 3.5h and permanent ischemia group (PO.05). The infarct areas in reperfusion groups after 1.5h at 24h significantly differed from those of reperfusion groups after 3.5h and permanent ischemia group (P<0.05). (5) In COX-2 immunohistochemical stain, the positive cells were present in the infarct periphery till 4 or 5 hours, with highest expression during 6 and 9 hours, but were absent in infarct core within 24 hours.Conculsion (1) With the development of infarct lesions, the range of IP decreased progressively from core to periphery. The time window of IP was related to the location within infarct lesions. The time window of IP adjacent to infarct core (3 hours) was present shorter than that of IP apart from infarct core (9 hours). (2) The sensitivity to assess IP by anisotropy was less than that by ADC and DCavg. (3) The region just adjacent to the infarct lesion was not IP and never evoluted to necrosis, demonstrating normal signal intensity on DWI, slightly limited diffusion and slightly elevated anisotropy. (4) The positive cells of COX-2 immunohistochemical stain expressed remarkably after 6 hours, suggesting that apoptosis occurred in amount of cells in the infarct periphery and IP has been transferred to irreversible necrosis tissue. (5) The time window of effective reperfusion should be lower than 2.5h.Part II: The imaging evolution of monkey brain ischemic penumbra Objective To study normal CT perfusion and MRI features of normal monkey brain and imaging evolution of IP in monkey MCAO model.Materials and Methods Seven male monkeys were used to make MCAO model by interventional methods. CT plain scan and perfusion examination were performed with GE LightSpeed pro 16 CT scanner. MR imaging was performed with GE 1.5T Twin Speed Infinity with Excite I magnetic resonance system. Head coil was used. The sequences included axial TiWI, T2WI, contrast-TiWI, DWI, DTI, PWI, MRSand MRA, which were performed before MCAO and lh, 5h, lOh, 15h, 20h, 24h and 48h after MCAO respectively. Parameters measured included cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), ADC, DCavg, FA, RA, 1-VR, negative enhancement integral (NEI), N-acetylaspartate (NAA), choline (Cho), creatine (Cr), lactate and lipid (LL), NAA/Cr and Cho/Cr. ROIs of normal monkey brain included cortex, medulla and basal ganglia. Four ROIs of infarct lesion were measured based on T2WI at 24 hours. Point 1, 2 and 3 were located within the high signal intensity region. Point 1 was infarct core, Point 3 was infarct margin, and Point 2 was the midpoint between Point 1 and 3. Point 4 demonstrated normal signal intensity just adjacent to high signal intensity. The relative ratios between infarct lesions and collateral normal brain were calculated. Statistical analysis was performed by SPSS 11.0 software package. Differences were considered statistically significant at P less than 0.05.Results (1) CBF, CBV, FA, 1-VR and NEI within different areas of normal monkey brain were significantly different (P<0.05). Those of medulla significantly differed from those of cortex and basal ganglia (PO.05). None of those were significantly different between cortex and basal ganglia (P<0.05). Cho, Cr, NAA and NAA/Cr were significantly different among different areas (P<0.05). (2) Five of seven monkey MCAO models were performed successfully. Viable imaging parameters of different areas within infarct lesions evoluted viably. Point 1, 2 and 3 were close to each other at about 15h. (3) There were were significantly different of rCBF and rNEI within 20h, of rCBV and rDCavg within 15h, of rADC and r(l-VR) within lOh, of rFA and rRA within 5h, of rMTT at 24h (P<0.05). (4) rCBV, rCBF, rMTT, rDCavg, rADC and rNEI of Point 4 significantly differed from those of Point 1, 2 and 3 within 24 hours (PO.05). rFA and rRA of of Point 4 significantly differed from those of Point 1, 2 and 3 at lh (PO.05). r(l-VR) of Point 4 significantlydiffered from those of Point 1, 2 and 3 within 10 hours (PO.05). (5) The probability equations of IP by Binary Logistic regression analysis were composed of rCBF, rADC, rDCavg, rNEI and occlusion time, respectively. The equations could be only composed of rCBV. The greastest force of equation was rCBF. (6) The thresholds of IP by ROC curve were 0.203 of rCBF, 0.483 of rCBV, 0.571 of rADC, 0.614 of rDCavg and 0.25 of rNEI. (7) The infarct areas of DWI, ADC, DCavg and MTT were no significantly different within 24h (P>0.05). Those of CBF, CBV and NEI were significantly different within 24h (PO.05). (8) The infarct areas at each time were no significantly different between DWI, ADC and DCavg, between NEI and CBV (P>0.05). Those of MTT significantly different from those of DWI, ADC, DCavg and CBF within 10 hours (P<0.05). Those of MTT significantly differed from those of NEI and CBV within 24 hours (PO.05). Those of CBF significantly differed from those of NEI and CBV after 10 hours (PO.05). (9) Design a picture-process computer software to demonstrate IP by thresholds. The IP areas calculated by this method significantly differed from PWI(MTT)-DWI mismatch region (PO.05). The former were larger than the latter before 15 hours, but smaller at 20h and 24h. 00) None of rCho/Cr were significantly different at any time (P>0.05). There were significantly different of rNAA/Cr within 24 hours, of rLL within 20 hours (PO.05). There were significantly different of rNAA/Cr between Point 1 and Point 3 within 15 hours, and of rLL at lh and 5h (PO.05). Point 4 significantly differed from Point 1 and 3 of rNAA/Cr within 24 hours and of rLL within 20 hours (PO.05).Conculsion (1) Because of different physiologic features, the parameters of blood dynamics, anisotropy and MRS metabolic materials were significantly different. (2) The imaging evolution of IP: With the development of ischemia, the necrosis tissue expanded from infarct core to infarct periphery. But the spacial evolution was not linear. IP adjacent to infarct core transferred shorter than IP apartfrom infarct core. The parameters of infarct core were not invariable. (3) The time window of IP should be divided into therapy time window and presence time window. They had different meanings. The presence time window of IP in monkey MCAO model was 15 or 20 hours. (4) CBV was the best marker to assess whether the tissue was viable. And CBF was the best marker to assess the presence of IP. (5) The best imaging examination methods for hyperacute and acute brain infarction were DWI first, and then CT or MRI perfusion examination. CT perfusion examination was better. (6) The thresholds of IP were 0.203 of rCBF, 0.483 of rCBV, 0.571 of rADC, 0.614 of rDCavg and 0.25 of rNEI. (7) IP areas: PWI-DWI mismatch regions could not estimate IP areas actually. In the early time after onset they would be smaller than the actual areas, but would be benign oligemia in the later time. CT perfusion (MTT) and DWI mismatch regions were much more close to the actual IP. IP thresholds combined with self-made computer software should demonstrate IP areas visually. IP accounted for 20~38% of infarct area at lh. It reduced to 15~35% at 5h~10h and 13—25 % at 15h. Then it declined rapidly, 9~ 15 % at 20h and 3 ~ 12 % at 24h. (8) There was limitation to assess IP by MRS. The value of NAA to assess IP was better than that of Lac, but the sensitivity of Lac to ischemia was much more higher than NAA.