| Part one: inter- and intra-observer agreement of ADC measurements of lung cancer in free breathing, breath-hold and respiratory triggered diffusion-weighted MR imagingObjective: To prospectively evaluate inter- and intra-observer agreement of apparent diffusion coefficient(ADC) measurements in free breathing, breath-hold and respiratory triggered diffusion-weighted MR imaging of lung cancer.Methods: Twenty two patients with lung cancer underwent free breathing, breath-hold and respiratory triggered diffusion-weighted imaging on 3.0T MRI. The ADC values of the lung cancer were calculated twice by both of the two independent observers and were compared using one-way analysis of variance(ANOVA). Inter- and intra-observer agreements of ADC were assessed by intraclass correlation coefficients(ICC) and Bland-Altman plot.Results: The mean ADCs in breath-holding, respiratory triggering and free breathing by observer 1 were(1.289±0.320)×10-3mm2/s,(1.292±0.270)×10-3mm2/s and(1.365±0.299)×10-3mm2/s, and those by observer 2 were(1.280±0.256)×10-3mm2/s,(1.300±0.271)×10-3mm2/s and(1.357±0.292) ×10-3mm2/s. No statistical significance was found amond the mean ADCs in different imaging methods(observer 1: F =0.909, P=0.406; observer 2: F =0.945, P=0.391). The results showed great intra- and inter-observer agreement(ICC: 0.913~0.998, WCV: 3.4%~5.7%).Conclusion: ADC measurements of lung cancer in free breathing, breath-hold and respiratory triggered diffusion-weighted MR imaging had good intra- and inter-observer agreement. Part two:Measurement reproducibility of maximum diameter derived by CT and MR imaging in lung cancerObjective:To investigate the reproducibility of MRI routine sequence(T1WI, T2 WI, DWI) and CT in peripheral lung cancer and central lung cancer, and to provide reference standard for evaluating the treatment response of lung cancer.Methods: 53 cases of patients with lung cancer were examined using SIEMENS’s 16 layer, 64 layer of MSCT and 3.0T Verio MR scanner. The longest diameter of lung tumor in the transverse images were measured on each image according to the RECIST1.1 standard, including CT lung window, CT enhanced mediastinal window(arterial phase), MRI T2-BLADE, axial T1-VIBE and DWIb0, DWIb300, DWIb800. The consistency analysis of the length diameter was assessed with the intraclass correlation coefficient(ICC), and the Bland-Altman method was used to evaluate the distribution of measurement points.The difference analysis were assessed by paired samples t test and nonparametric rank sum test, P<0.05 was considered statistically significant.Results: 33 cases of peripheral lung cancer and 20 cases of central lung cancer, both unenhanced lung window and enhanced mediastinal window, the consistency of tumors’ length derived by CT and the conventional MRI sequences(T1-VIBE, T2-BLADE, DWIb0, DWIb300, DWIb800) was good(ICC>0.75). There was no significant difference in the length between CT unenhanced lung window or enhanced mediastinal window and T1-VIBE, T2-BLADE, DWIb300, DWIb800, DWIb0 in the peripheral lung cancer. As for central lung cancer, the tumor length derived from enhanced CT mediastinal window was different from MRI sequences(T1-VIBE, T2-BLADE, DWIb0, DWIb300, DWIb800), but the length derived by T1-VIBE, T2-BLADE was not significantly different from DWI sequences(b=0, 300 s/mm2, 800 s/mm2).Conclusion: For peripheral lung cancer, CT and routine MRI sequences(T1WI, T2 WI, DWI) have agreement with morphology measurement, and they can replace by each other; but for central lung cancer, there are differences between enhanced CT and conventional MRI sequences(T1WI, T2 WI, DWI), that is, the alternate use in evaluating the treatment response of lung cancer should be paid attention; the length measurement of DWI in lung cancer is consistent with that of T1 WI and T2 WI, suggesting that DWI can not only provide functional information but also provide quite reliable morphological information.Part three: Correlative Study between CT Perfusion(CTP) Parameters and Apparent Diffusion Coefficient(ADC) of non small cell lung cancer(NSCLC)Objective: To investigate the correlation between CTP parameters and ADC of NSCLC, to provide a basis for whether CTP and diffusion weighted imaging(DWI) can be used as a supplement in the evaluation of therapeutic efficacy.Methods: 23 lung cancer patients(23 lesions) underwent CTP and DWI examination between February 2012 and August 2014 were included in this study. CTP scan used 64 slice spiral CT scanner. The scanning started after the injection of contrast agent and the scanning time was 45 s. Reconstruction layer thickness was 4 * 7.2mm, matrix was 512 * 512 and Z axis coverage was 2.88 cm. Ultravist(370mg I/ml) 50 ml was bolus through elbow vein by a high pressure syringe and the injection rate was 5.0~6.0ml/s. The workstation(MMWP, Siemens) could automatically generate MIP and a series of perfusion maps, including blood flow(BF), blood volume(BV), Patlak blood volume(PBV), time to start(TTS), time to peak(TTP), permeability surface(PMB), mean transport time(MTT) etc.. The fixed area of region of interest(ROI) and the maximum area of ROI were drew at the largest crosss-section of tumor. DWI scan used Siemens 3.0T Verio MR instrument and the axial single shot spin echo planar diffusion weighted imaging was obtained. Three b values(b=0, 300s/mm2, 800s/mm2) were selected. The thickness was 5 mm and the interval was 6 mm. In the post processing workstation(VE40A Syngo, Siemens) the ADC map(b=0, 300s/mm2), ADC map(b=0, 800s/mm2) and ADC map(b=300s/mm2, 800s/mm2) were generated. The selection of ROI was as same as above. Perfusion parameters and ADC from two ROIs were compared by paired sample t test or nonparametric Wilcoxon test. Under the same ROI, pearson correlation analysis was used to analyse the correlation between CTP parameters and ADC values obtained by different b values. r≥0.7 meaned highly correlated, 0.4≤r<0.7 meaned generally correlated and r<0.4 meaned poorly correlated. P<0.05 had statistically significant difference.Results: Among all the perfusion parameters in this study, only BF, BV and MTT would be different because of ROI’s morphology. However, BV(P), TTS, TTP and PMB(P) were independent of the morphology of ROI. No matter which b value to obtain ADC, the ADC values obtained from two ROIs always had statistical differences. In addition, under the same ROI, the ADC obtained from different b values also had the difference. Among all the perfusion parameters, only under ROIf did PMB(P) with ADC(b=0, 800s/mm2) and ADC(b=300s/mm2, 800s/mm2) have a general correlation, which was showed positive linear relationship(r=0.433, P=0.039; r=0.537, P=0.008), however, the rest perfusion parameters and ADC values were not correlated.Conclusion: The correlation between CTP parameters and ADC of lung cancer was poor, so they couldn’t be replaced by each other. In the application of lung cancer, CTP and DWI should be complementary to each other, and give full play to their respective value.Part four: Value of diffusion-weighted imaging and computed tomography perfusion in predicting the efficacy of non small cell lung cancer chemoradiotherapyObjectives: To assess the clinical value of CTP parameters and DWI parameters in predicting the efficacy of NSCLC chemoradiotherapy in terms of two-year progression-free survival(PFS) and overall survival(OS).Methods: 23 patients(23 lesions) pathologically diagnosed with non small cell lung cancer were reviewed, and all underwent CTP and DWI scans prior to treatment in our hospital between February 2012 and August 2014. CTP was performed on a 64-slice CT scanner, with reconstructed slice thickness of 4×7.2mm, matrix of 512×512 pixels and x-axis of 2.88 cm. Immediately after a bolus injection of 50 ml Ultravist(370mg I/ml) at 5.0~6.0ml/s via median cubital veins using a high pressure injector, the scan started and lasted 45 s. The workstation(MMWP, Siemens) automatically produced MIP and a series of pseudo-color perfusion images, including BF(blood flow), BV(blood volume), PBV(patlak blood volume), TTS(time to start), TTP(time to peak), PMB(permeability), MTT(Mean Transit Time), etc. The tumor slices were found and both fixed area of ROI and maximum area of ROI were drawn. Siemens 3.0T Verio MR machine was used to perform the axial single-shot spin-echo echo planar diffusion-weighted imaging. DWI images were acquired with b values of 0, 300s/mm2 and 800s/mm2, slice thickness of 5 mm and inter-slice distance of 6 mm. For post-processing the data were transferred to a Siemens Syngo VE40 A workstation, and the same ROIs as before were chosen to generate ADC images when b=0, 300s/mm2, b=0, 800s/mm2, and b=300s/mm2, 800s/mm2 respectively. The patients were treated with concurrent chemoradiotherapy and the efficacy was assessed in terms of PFS and OS according to the RECIST 1.1 standards. Kaplan–Meier analysis was applied for PFS and OS survival curves and log-rank test for comparisons among different factors and the optimal threshold was identified. Cox regression with single variable was used to analyze various predictors for rates of two-year PFS and OS and Cox regression with multiple variables to identify the independent predictors, significant level being 0.05.Results: Clinical stages: 4cases of IIB(17.4%), 6 cases of IIIA(26.1%), 9 cases of IIIB(39.1%) and 4 cases of IV(17.4%). The rates of two-year PFS and OS were respectively 30.4 %(7/23) and 39.1 %(9/23). No significant correlation was found between the clinical stages and the rates of two-year PFS and OS(P=0.223 and 0.112). Regression analysis with multiple variables showed that CTP parameters of independent predictive value for PFS and OS were BFm(P=0.002, HR=3.98), PMB(P) m(P=0.001, HR=4.12), and that ADCs of independent predictive value for PFS was ADC2m(P=0.001, HR=4.53) and to OS were ADC2m(P=0.005, HR=3.99), ADC3m(P=0.027, HR=2.44). A prediction scoring model was established accordingly: 0 for those which had BFm≥46.10 ml/100ml/min, PMB(P)m≥30.95 ml/100ml/min or ADC2m≤1.52×10-3mm2/s, 1 for the otherwise. Each case was scored 3 at most and 0 at least, resulting in three groups: 0/1 score, 2 scores and 3 scores, of which the median PFS were 26, 17 and 11 months, the median OS were 29, 21 and 12 months, while the two-year PFS rates were 71.4%, 18.2% and 0.0%, the two-year OS rates were 85.7%, 27.8% and 0.0%, respectively. The differences between groups had statistical significance(P<0.001).Conclusions: Clinical stages have limited values in predicting the efficacy of chemoradiotherapy for NSCLC. BF, PMB(P) and ADC(b=0, 800 s/mm2) had independent predictive value for PFS and OS, and the prediction scoring model on the basis of which had higher predictive value for treatment efficacy. |
PDF Full Text Request