【作者】 胡晓云；

【导师】 胡春洪；

【作者基本信息】 苏州大学 ， 放射医学， 2014， 博士

【摘要】 第一部分双源CT肺部灌注成像及其检测放射性肺损伤的临床意义目的：针对胸部放疗后的放射性损伤（radiation-induced lung injury, RILI）常规影像形态学改变出现较晚、临床难以早期诊断的问题，胸部功能成像能否比传统影像更早期反映RILI的功能变化？本研究探讨应用双源CT肺部灌注成像（CTperfusion imaging, CTPI）技术的可行性及其早期诊断RILI的临床价值。方法：选取临床筛查肺动脉栓塞而行4D-CT增强检查但影像诊断结果阴性以及临床最终排除了肺部疾患的患者20例作为对照组，年龄47-76岁（男、女各10例）。实验组为48例接受术后放疗的上段食管癌或胸腺肿瘤患者(照射总剂量均为60Gy)，年龄43-70岁（男27例，女21例）。放疗前及放疗1/2总剂量时间点(30Gy)行CTPI检查，同期检测外周血中肿瘤坏死因子(TNF-α)、转化生长因子(TGF-β1)，分析发生RILI(A组)与未发生RILI(B组)患者的血清细胞因子、常规CT表现及CTPI灌注值[相对血流量(rrBF)、相对血容量(rrBV)、相对毛细血管通透性(rrPS)]的变化，采用随机区组设计t检验比较两组间血清细胞因子、CTPI灌注值的差异，采用χ2检验比较常规CT与CTPI对RILI检出的差异。使用西门子新双源CT机（FLASH）肺部4D容积扫描技术作为CTP成像方法，先行常规高分辨率CT平扫（HRCT），后行灌注成像。结果：⑴对照组所有20例患者均能获得层次丰富、结构清晰的CTP图像，肺野内任意取不同大小的感兴趣区（ROI），均能获得重复性好的灌注值；正常成人肺的rBF、rBV及rPS平均值分别为149.3±18.3mL/100mL/min、14.86±2.65mL/100mL、9.54±2.91mL/100mL/min；肺部各灌注参数及密度值（HU）男女性别、左右部位差异均无统计学意义（P>0.05）；上、下肺野ROI的rBF、rBV值有显著差异（P<0.05），上中肺野、中下肺野之间的各参数值无明显差异（P>0.05）。⑵48例患者中，18例发生RILl(A组)。A组外周血TNF-α和TGF-β1放疗前后的差异无统计学意义(均P>0.05)。放疗1/2总剂量时，A、B两组的外周血TNF-α、TGF-β1组间差异无统计学意义(均P>0.05)；但A组常规CT图像上有2例出现阳性征象。A组受照射肺组织rrBF、rrBV、rrPS均较照射前显著增高(均P<0.05)；B组的rrBF、rrBV较照射前有增高，差异有统计学意义(均P<0.05)，rrPS无明显变化(P>0.05)；照射后A、B两组rrBF、rrBV、rrPS间差异均有统计学意义（均P<0.05）。根据ROC曲线，设rrPs=1.22为阈值，诊断RILl的敏感度、特异度分别为88.9％、90.0％，优于HRCT的11.1％、90.0％(χ2=13.61,P<0.05)。⑶对照组与实验组放射治疗前的各对应灌注值之间差异无统计学意义（P>0.05）。结论：⑴双源CTPI技术能够用于肺组织的灌注成像研究，能定量分析肺组织血流动力学参数（rBF、rBV、rPS等）。⑵外周血中TNF-α、TGF-β1变化对早期检测肿瘤放疗患者RILI的价值尚不确定。⑶CTPI能反映放疗后肺组织血液动力学的变化，能早期反映RILI患者照射野的异常灌注，有可能成为早期检出RILI的有效工具。第二部分放射性肺损伤兔阶梯模型的制备方法及鉴定一、放射性肺损伤（RILI）兔阶梯模型的制备目的：针对人类放射性肺损伤（RILI）各期难以在临床上准确把握以及难以进行生化、影像、组织病理等重复性研究，故建立一种适合动态观测与研究的动物模型非常必要。本研究旨在探讨放射性肺损伤（RILI）兔阶梯模型的制备方法。方法：健康新西兰大白兔54只，随机分作A组、B组及C组，每组各18只。三组均由同一麻醉师使用同样的混合法麻醉技术，即先肌注地西泮（用量为0.8~1.0mL/kg），后肌注速眠新II号（用量为0.4~0.5mL/kg），然后经自制中型动物医学成像扫描床固定后作高能X线单侧全肺单次照射，照射剂量分别35Gy、25Gy及15Gy，分别于照射后第1、2、3、4、5、6、8、12、16、20、24周进行阶梯性常规CT观察。重点观察混合麻醉法的安全及有效性，以及比较各模型组出现CT阳性表现的时间及各组模型的总死亡率。结果：⑴三组模型麻醉效果显著，无意识的麻醉时间平均为30-40分钟，麻醉死亡率均为零。⑵A、B、C三组模型常规CT出现渗出性毛玻璃影的高峰时间点分别为2周、4周、16周，并均在随后的阶梯时间点得到进一步的证实，即RILI的产生；模型中途死亡数分别为4只、1只、1只，死亡率分别为22.2%、5.6%、5.6%。结论：混合法麻醉技术具有安全性高、重复性好、可控性强、易于操作等优点。B组（25Gy）具有成模时间适中、死亡率低、阶梯匀称等优势，利于进一步的RILI阶梯性分析，是首选的造模方法。二、放射性肺损伤阶梯模型的鉴定目的：临床上放射性肺损伤（RILI）的发生、发展的演变过程较为复杂，RILI的各期无明显分界，因此对RILI动物模型的动态观测及鉴定具有重要的临床指导意义。故本研究旨在探讨RILI兔阶梯模型的动态血清学及组织病理学鉴定，以确定阶梯模型的发生及演变过程。方法：麻醉方法同上，采用上述25Gy造模方法，共计60只健康新西兰大白兔经高能X线单侧全肺单次照射造模，设为实验组。同时随机同批次大白兔36只作假照射，设为为对照组，按下述照射后时间点分12个亚组，实验及对照组的每个亚组分别为5只、3只。分别于照射后第1、6、12、24、48、72小时及第1、2、4、8、16、24周进行开胸暴露心脏后穿刺提取4-8mL全血以备血清学检测，随后处死模型分别取两肺中带上、中、下野6处标本，分别进行HE染色光镜、电镜和局部肺组织TNF-a、TGF-β1的检测。实验组各时间点若有模型死亡，及时添加新个体进行补充。结果：⑴实验组两只分别于照射后第2、8周死亡，实验周期内总死亡率3.3%。⑵实验组所有兔受照射肺均产生了RILI，早期以急性炎症反应为主，晚期以进行性肺纤维化为特征。⑶实验组受照射1小时后局部肺组织TNF-α表达、48小时后TGF-β1表达与对照组差异有统计学意义（P值均＜0.05）。⑷光镜下，实验组受照射1小时后肺泡壁厚度、肺间质面积密度、24小时后间质内纤维母细胞和纤维细胞数量与对照组差异均有统计学意义（P值均＜0.05），并分别与照射后的时间直线相关（r=0.82086、0.87181、0.68230，P值均<0.05）。⑸电镜下，实验组各时间点之间胶原纤维相对含量差异有统计学意义（F=100.31，P=0.000），对照组各时间点之间的差异无统计学意义（F=1.00，P=0.450）。实验组受照射48小时后肺内胶原纤维相对含量与对照组差异有统计学意义（P值均＜0.05），并与照射后时间直线相关（r=0.99318，P=0.0000）。结论：RILI兔阶梯模型具有良好的可靠性、稳定性以及RILI发生发展的阶梯性，较好地模拟了RILI的发生、发展的演变过程。细胞因子等对RILI的诊断有一定的参考意义，但不能作为诊断或预测的依据。第三部分放射性肺损伤兔阶梯模型功能成像技术研究目的：探讨健康新西兰大白兔的CT灌注成像（CTPI）技术参数及其定量分析的可行性。方法：健康新西兰大白兔36只假照射组（雌雄各18只），均由同一麻醉师使用同样的混合法麻醉技术，即先肌注地西泮（用量为0.8~1.0mL/kg），后肌注速眠新II号（用量为0.4~0.5mL/kg），然后经自制中型动物医学成像扫描床固定后，分别于假照射后照射后第1、6、12、24、48、72小时及第1、2、4、8、16、24周进行阶梯性常规CT扫描及CT灌注成像（CTPI）。CTPI使用设备为西门子新双源CT机（FLASH），采用全肺4D容积动态扫描技术。扫描参数：管电压100kv，管电流自动调节；4D1.5s模式；非离子对比剂规格为300mg I/mL，总量5mL，注射流速0.8mL/s。CTPI时采用专用动物呼吸机进行呼吸暂停15s。结果:36只大白兔均能获得层次丰富、结构清晰的CTPI图像，肺野内任意取不同大小的感兴趣区（ROI），均能获得重复性好的灌注值；正常兔肺的rBF、rBV及rPS平均值分别为123.8±25.9mL/100mL/min、13.04±2.07mL/100mL、8.97±2.14mL/100mL/min；肺部各灌注参数及密度值（HU）在雌雄性别、左右部位等方面的差异无明显统计学意义（P>0.05）；上、下肺野ROI的rBF、rBV值的差异有显著统计学意义（P<0.05）。假照射后不同时间点的各对应灌注值重复性佳，各时间点CTPI参数差异无明显统计学意义（P>0.05）。结论:正常兔全肺CTPI各参数值能够进行定量测量与分析，且重复性好，与人类全肺具有相似的变化规律。第四部分双源CT灌注成像早期诊断放射性肺损伤的实验研究目的：利用全肺双源CT灌注成像（CTPI）技术研究放射性肺损伤（RILI）兔阶梯模型的血流动力学变化规律，探索其发病进程中的病理基础及其早期诊断RILI的价值。方法：健康新西兰大白兔72只，利用随机数字表分为两组：实验组36只行25Gy单侧全肺单次照射，对照组36只为空白对照组，行单侧全肺假照射；两组按照射后1、6、12、24、48、72h和1、2、4、8、16、24周各分为12个亚组。每只兔于照射前和照射后所处亚组时间点做常规CT和CTPI，后取肺组织行光镜、电镜观察，将同期CTPI表现、病理变化及HRCT征象作对照分析。实验组与对照组CTPI参数比较采用t检验，CTPI参数与病理观察值的相关性采用两变量直线相关分析，CTPI与常规CT对RILI检出率比较采用卡方检验。结果：⑴对照组肺CTPI参数在各时间点上保持在相对稳定的水平。实验组照射后肺实质血流量（rBF）、血容量（rBV）和毛细血管通透性（rPS）在多数时间点上（6、12、72h及1、2、8、16、24周）与照射前差异有显著统计学意义（t=2.90～6.37，P＜0.05），呈“先降—后升—再降”的规律：1～12h先短暂的降低，后快速升高并显著高于对照组，于72h至2周分别达高峰，随后下降，并逐步低于对照组，至24周达最低值。⑵实验组受照射肺病理变化以毛细血管内皮细胞、基底膜和肺泡上皮细胞损伤为主，并随时间呈现一定规律性。实验组照射后肺rBF、rBV与病理改变有显著相关性（r=0.74、0.83，均有P＜0.05），rPS与毛细血管外红细胞数、毛细血管基底膜破坏之间有显著依存关系（r=0.87、0.88，均有P＜0.05）。⑶在所有时间点上，CTPI对RILI的检出率均高于HRCT，两者差异有统计学意义（χ2=4.37，P=0.036）。根据ROC曲线，两者诊断RILI的敏感度、特异度分别达92.3%、90.0%与69.2%、90.0%（Z＝13.06，P<0.05），CTPI的诊断效能明显优于常规CT，尤其是在RILI模型的早期阶段（放疗后的前4周）。结论：CTPI参数初步揭示了RILI过程中的血流动力学演变规律，并间接反映以肺毛细血管通透性为主要变化的病理生理状态；RILI的CTPI异常在时间上先于常规CT，有望成为早期检测与早期诊断RILI的有力手段。更多还原

【Abstract】 PartⅠ clinical significance of early detection of radiation-induced lunginjury with dual-source CT perfusion imagingObjectives: It’s hard to make an early diagnosis due to the delayed morphologicalchanges using conventional imaging in the detection of radiation-induced lung injury(RILI) after thoracic radiotherapy. As to the issue, can thoracic functional imagingdeliver the earlier functional changes of RILI than traditional imaging? This studyapplies pulmonary perfusion imaging (CT perfusion imaging, CTPI) with dual-sourceCT (DSCT) scanner to explore the feasibility of the early diagnosis of RILI and itsclinical value.Methods: As the Control Group,20cases, age47-76(10males,10females) wereselected from the negative patients who were suspected with pulmonary embolism ruledout by the results of4D-DSCT enhancement imaging and clinical diagnosis.48patients,aged43-70(27males and21females) with postoperative radiotherapy with upperesophageal cancer or thymic tumor (irradiation dose of total, ie DT,60Gy) wereselected as Test Group. CTPI with detection of tumor necrosis factor (TNF alpha) andtransforming growth factor (TGF-beta1) in peripheral blood, were performed beforeradiotherapy and at1/2DT time point (30Gy). Serum levels of cytokines, conventionalCT manifestations and CTPI perfusion values [relative blood flow (rrBF), relative bloodvolume (rrBV), relative blood capillary permeability surface (rrPS)] were compared andanalized between patients with RILI (group A) and non-RILI (group B). Using Siemensnew dual-source CT machine (FLASH)4-D lung volume scanning technology as CTP imaging method, routine high resolution CT scan (HRCT) was first performed thenperfusion imaging.Results:⑴In all subjects of Control Group, the satisfactory CTP images wereproduced with nice layers and structures. Stable data could be gained from eachdifferent size of ROIs in lung field; The normal pulmonary BF, BV and PS were149.3±18.3mL/100mL/min,14.86±2.65mL/100mL,9.54±2.91mL/100mL/min,respectively; there was no significant difference between the genders or the bilateralsides for all the above parameters obtained (P>0.05); there was significant difference,for BF and BV, between the superior and inferior lung fields（P <0.05）and nosignificant difference for all perfusion parmeters, between superior and middle ormiddle and inferior fields (P>0.05).⑵18cases (18/48) of ARILI were diagnosed. Thediffereces of TNF-α and TGF-β1pre-and post-radiotherapy were not significant inGroup A(P>0.05), but there was positive sign of conventional CT in two cases. Themean values of rrBF, rrBV and rrPS (relative regional blood flow, volume andpermeability surface, respectively) of post-radiation in Group A were correspondinglyhigher than those of pre-radiation(each P <0.05); the rrBF and rrBV of post-radiationfrom Group B were significantly higher than those of pre-radiation (each P<0.05), butno obviously change for rrPS (P>0.05). There were significant differences for allperfusion values detected pre-and post-radiation from radiation lung fields betweenGroup A and B (each P <0.05). Applying the diagnostic threshold value of rrPS=1.22,the sensitivity, specificity, positive and negative predictive values of CTPI (88.9%,90.0%) for early diagnosis of ARILI were better than those of CT (11.1%,90.0%;χ2=13.61, P <0.05).⑶There were no significant differences for all correspondingperfusion values of pre-radiation between control and test groups.Conclusions:⑴Dual-source CTPI may apply in the study of pulmonary perfusionimaging via providing the quantitative analysis of pulmonary hemodynamic parameters,such as rBF, rBV, rPS, etc.⑵TNF-α, TGF-β1of peripheral blood may not yield apositive value for early diagnosis of RILI in patients with radiotherapy.⑶Values ofCTPI parameters may reflect the hemodynamic changes of post-irradiation lung,yielding a promising value in early diagnosing ARILI. PartⅡ Establishment and appreciation of ladder model ofradiation-induced lung injury in rabbits1. Preparation method of the ladder model of RILI in rabbitsObjective: For human RILI, it is difficult to accurately control each phase of RILIand to perform repeatable study of biochemistry, imaging and histopathology. Therefore,to establish a suitable for dynamic observation and research of animal model is verynecessary.This study was to explore the preparation method of the ladder model of RILIin rabbits.Methods: Fifty-four healthy New Zealand rabbits were randomly divided intothree groups: A, B, and C (18rabbits for each group). the same hybrid methodanesthetic technique was used by the same anesthetist in the three groups, that is, thefirst muscle injection of diazepam (the dosage is0.8-1.0mL/kg), then intramuscularinjection of Su-Mian-Xin II (the dosage is0.4-0.5mL/kg), and then underwenthigh-energy X-ray irradiation in unilateral single lung after fixed by the homemademedium animal medical imaging scanning bed. The irradiation dose of total was35,25and15Gy, respectively. Conventional CT observation was performed at ladder timepoints (1,2,3,4,5,6,8,12,16,20,24weeks) of post-radiation, including the safetyand efficacy of hybrid anesthesia method, comparison of time of CT positive expressionand total mortality of each model.Results:⑴The anesthesia efficacy in three groups of model were all remarkablewith unconscious time of anesthesia for an average ranging from30to40minutes, andzero anesthesia mortality.⑵exudative ground-glass appearance of peak time inconventional CT was2w,4w,16w, respectively for A, B, C groups, which wereconfirmed by further study in the following ladder time points, namely the generation ofRILI. Model deaths during the test were4,1,1, respectively, that is, the correspondingmortality rates were22.2%,5.6%and5.6%, respectively.Conclusions: Hybrid method anesthetic technique may provide such advantages ashigh security, good repeatability, strong controllability, easy operation, etc. Group B (25 Gy) with the advantages of moderate model-building time, lower mortality rate, theladder symmetry, facilitates further RILI analysis at ladder time points, and is thepreferred method of model building.2. Appreciation of ladder model of radiation-induced lung injury in rabbitsObjective: In clinic, the evolution process of occurrence and development of RILIis relatively complex, and the periods or phases of RILI have no obvious dividing line.So the dynamic observation and identification of animal models of RILI were yieldingan important clinical significance.This study was to explore dynamic serological andhistopathologic evaluation in rabbit model of RILI in order to determine the occurrenceand evolution process at ladder time points.Method: Using the above25Gy model-building method and the ditto anesthesiamethod, a total of60healthy New Zealand white rabbit, set as the experimental group,underwent unilateral single lung irradiation with high-energy X-ray exposure. At thesame time, as for the control group, the random batch of36white rabbits performedwith a sham-irradiation. According to the following time points of post-irradiation,12subgroups were divided and5and3rabbits for each subgroup in experiment and controlgroup, respectively. Extraction of4to8mL whole blood for serological detection viaheart puncture after thoracic open at the such ladder time points as1,6,12,24,48,72hours and1,2,4,8,16,24weeks after exposure, then6specimens were taken fromsuperior, middle and inferior field of bilateral lungs. Afterwards, the pathologicalchanges were observed with light and electron microscopies; the expression of TNF-aand TGF-β1in local lung tissue was detected by immunohistochemistry. During theobservation of models, a new individual was added to supplement if there was modeldeath.Result:⑴In experimental group, two died at2w,8w after irradiation,respectively, the total mortality in the experimental cycle was3.3%.⑵RILI occured inall rabbits of test group: at early stage, it’s characterized by acute inflammatory reaction;it featured the progressing fibrosis at later stage.⑶Regarding expression of TNF-a andTGF-β1after1st and48th of postirradiation, respectively, there were statistically difference between test and control groups (each P <0.05).⑷In terms of the aspects ofthickness of alveolar wall, density of pulmonary interstitium (after1st ofpostirradiation), amount of fibroblast and fibrocyte from interstitium (after24th ofpostirradiation), there were statistically difference between two groups (each P <0.05),and test group correlated well with the time of post-irradiation (r=0.82086,0.87181,0.68230; each P <0.05).⑸There was statistically difference among the relative amountof collagen fibers at time points of post-irradiation in test group (F=100.31, P=0.000),while no difference in control group (F=1.00, P=0.450). As to the relative amount ofcollagen fibers after48th of postirradiation, test group differs from control groupsstatistically (each P <0.05), correlating well with the time of post-irradiation(r=0.99318, P=0.0000).Conclusion: Stable and reliable ladder model of RILI in rabbits was made throughsingle fraction irradiation in whole unilateral lung with25Gy of high-energy X-rays,which may simulate the occurrence and development of evolution of RILI. Cytokinesmay be taken as a certain reference, but cannot be used as the basis for diagnosis orprediction.Part Ⅲ The study of technique of functional imaging in ladder modelof RILI in rabbitsObjective: To explore the feasibility of the technological parameters andquantitative analysis of CT perfusion imaging (CTPI) in healthy New Zealand whiterabbits.Method: Thirty-six healthy New Zealand rabbits (18, each gender) wereperformed with the same hybrid method anesthetic technique by the same anesthetist,that is, the first muscle injection of diazepam (the dosage is0.8-1.0mL/kg), thenintramuscular injection of Su-Mian-Xin II (the dosage is0.4-0.5mL/kg), and thenunderwent sham-irradiation in unilateral single lung after fixed by the homemademedium animal medical imaging scanning bed. Conventional CT (HRCT) and CTperfusion imaging (CTPI) were performed at ladder time points (1,6,12,24,48,72hours and1,2,4,8,12,16,24weeks) after sham-radiation. CTPI were performed by theequipment of Siemens new dual-source CT machine (FLASH), using the whole lung4-D volume dynamic scanning technology. Scan parameters: the tube voltage100kv,automatically adjust the tube current;4-D1.5s mode; total5mL of non-ionic contrastmedium (300mg I/mL), with injection rate0.8mL/s. Special animal breathing machinewas adopted for with15s apnea during CTPI.Result: In all subjects, the satisfactory CT perfusion images were produced withnice layers and well structures. Stable data could be measured from each different sizeof ROIs in lung fields. For healthy adult rabbits, the normal pulmonary BF, BV and PSwere23.8±25.9mL/100mL/min,13.04±2.07mL/100mL,8.97±2.14mL/100mL/min,respectively; there was no significant difference between the genders or the bilateralsides for all the above obtained parameters (P>0.05); there was significant difference,for BF and BV, between the superior and inferior lung fields （P <0.05）. Aftersham-radiation, each corresponding perfusion values at different time points deliver agood repeatability, and the CTPI parameters at each time point had no obviousdifferences with no statistical significance (each P>0.05).Conclusion: For normal adult rabbits, CTPI parameters in lungs can bequantitatively measured and analized with a good repeatability, and the similar changerule to that of the human lungs.Part Ⅳ Experimental study of early diagnosis for RILI withdual-source CT perfusion imagingObjective: To study of the hemodynamic change rules of ladder model of RILI inrabbit using whole lung CT perfusion imaging (CTPI) in order to explore the value ofearly diagnosis of RILI and its pathological basis in the development and process of thedisease.Method: Seventy-two healthy New Zealand rabbits were randomly divided intotwo groups:36rabbits in test group were administered with25Gy of single fractionatedirradiation in the whole unilateral lung; the other36rabbits in control group were sham-irradiated. According to the post-radiation time points (1,6,12,24,48,72h and1,2,4,8,16,24w, respectively), the test and control groups were divided into twelvesub-groups, respectively. All rabbits were respectively performed with conventional CT,CTPI and ultimately sacrificed at the different time points (1,6,12,24,48,72h, and1,2,4,8,16,24w) after irradiation. The pathological changes were observed with lightand electron microscopies; the expression of TNF-a and TGF-β1was detected byimmunohistochemistry. t-test, chi-square test and linear correlation analysis wereapplied in the comparison of results of CTPI, pathologic observed values between testand control groups.Result:⑴The changes of CTPI parameters from control group after irradiationwere relatively stable, but those of test group demonstrated a pattern ofdropping-rising-dropping in rBF, rBVand rPS with a significant difference from thoseof pre-radiation (t=2.90-6.37, each P <0.05): first, temporally dropping at1to12hours after irradiation, then strikingly rising (much higher than those of control group)to the peak at72hours to2weeks; lastly gradually falling down to the nadir (muchlower than those of control group) at24weeks.⑵In test group, capillary endothelialcells, basement membrane and alveolar epithelial cells, as the main injured targets,showed certain alterations in pathology. There was a significant correlation between thechanges of CTPI parameters (rBF and rBV) and pathohistology in control group (r=0.74,0.83; P=0.006,0.000), with the dependent relationship between rPS and the amounts ofRBC outside the capillary and the destruction of basement membrane (r=0.87,0.88; P=0.000,0.000).⑶The detection rate of RILI with CTPI was obviously higher than thatwith HRCT (χ2=4.37; P=0.036). According to ROC curve, the sensitivity andspecificity of CTPI and conventional CT for diagnosis of RILI were92.3%,92.3%and69.2%,90.0%, respectively (Z=13.06, P <0.05), denoting that CTPI diagnosticefficiency is superior to conventional CT, especially in the early stages of RILI model(4weeks after radiotherapy).Conclusion: CTPI parameters are able to reveal the rule of hemodynamic processand reflect the pathophysiologic state of different stages of RILI. By the time ofdetecting RILI, the detection rate of CTPI was clearly superior to that of HRCT, which yields potential value in predicting RILI.更多还原