不同年龄段人群腰椎CT值的快速求解及临床意义

发布时间：2018-05-21 20:45

本文选题：三维重建 + 骨密度　；参考：《南方医科大学》2014年硕士论文

【摘要】：背景在临床工作中发现大量腰腿痛的患者,有些患者甚至非常年轻,以男性多见。但是临床医生对于这些患者在进行体格检查时往往只是进行一些常规的检查,在发现某些具有阳性意义的体征时就让病人去做X线或者CT甚至MRI等影像学检查,得出的往往是腰椎椎间盘膨出、突出等诊断。直到患者出现了典型的症状甚至出现了骨折等严重的并发症的时候才得到骨质疏松症的确切诊断。这个时候既增加了患者的生活负担,降低了其生活质量,又增加了病人的经济负担。通过运用新的图像分割方法对这些病人的腰椎CT图像逐一分离出每一个椎体并且读取椎体模型的CT值最大值、最小值以及平均值后发现,一部分病人的椎体的CT值最小值均为负值,最低的时候甚至可以达到-1024。我们推测出这些椎体里应该含有气体的成分,并试着找寻椎体里会含有气体的原因。通常情况下老年患者会出现不同程度的腰椎退行性改变,椎体内可能存在气体,但是年轻的患者腰椎并不存在退行性病变,因此我们推测出年轻患者腰椎可能有其它一些病变,进一步调查发现患者的亲属存在原发性骨质疏松症病史。由此我们推测该患者可能具有原发性骨质疏松症的某些症状如骨密度值减低。目的将不同年龄段男女腰腿痛患者腰椎螺旋CT扫描图像导入三维重建软件,通过新的图像分割方法将这些图像中的腰椎椎体结构逐一分割出来迅速求解出腰椎CT值最大值、最小值以及平均值并探索其变化规律,并且尝试探讨CT值与骨密度值的关系。以达到对年轻人群进行筛查、对那些症状不典型的人群早期诊断、早期治疗,减少病人的经济负担和生活负担,降低其致残率以及并发症发生,提高其生存质量的目的。方法随机采集21-70岁202例腰腿痛病人,按照年龄以每隔10岁为一个年龄段进行分组,每个年龄段又按照性别再分别分为男女两组共10组,其中21-30岁年龄段男女各15例,31-40岁年龄段男女各22例,41-50岁年龄段男女各21例,51-60岁年龄段男女各21例,61-70岁年龄段男女各22例,每组男性女性各50块椎体,共500块。对病人进行腰椎螺旋CT扫描,CT设备为美国通用电气公司64排螺旋CT扫描仪。CT扫描的参数为120KV、264MAS,层厚为0.625mm,图像以512x512像素DICOM格式刻录至DVD光盘输出。首先采用新的图像分割方法对21-70岁202例腰腿痛病人共500块椎体通过三维重建软件将其逐一分离出来,其次求解出每一个年龄段男性和女性患者的腰椎CT值的最大值、最小值以及平均值并分别进行统计学分析。在Windows7Ultimate64bit操作系统下,将CT扫描数据导入医学三维重建软件Mimics14.0。随后重建出腰椎三维模型,先移除髂骨和骶骨,之后分别将各个椎体分离并建立相应的独立的蒙板(Mask)文件,再分离椎体、椎弓根及椎板围绕行成的闭合空间,接下来进行模型轮廓线的计算、填充Mask中的空洞、运用布尔运算中的“布尔减”操作将每个病人的腰椎椎体逐一分割出来。完成后读取椎体模型的CT值平均值、最大值及最小值。以SPSS16.0软件对实验数据进行统计学处理,采用单因素方差分析统计方法,对每个年龄段男女腰腿痛患者的CT值的最大值、最小值以及平均值进行比较。当读取的男性和女性不同年龄段组的CT值采用平均值±标准差表示时我们还比较相同年龄段男女性腰椎CT值平均值以及不同年龄段男性以及女性腰椎CT值平均值。分别对男性和女性不同年龄组之间患者CT值的最大值、最小值以及平均值进行单因素方差分析检验,看其差异是否具有统计学意义。若差异有统计学意义时再进行组间的多重比较。进行组间的多重比较时先对各部分数据进行方差齐性检验,方差齐时采用LSD法、Bonferroni法、SNK法以及Scheffe法。方差不齐时采用Dunnett's T3法和Dunnett's C法,P0.05认为差异有统计学意义。结果男性各组CT值平均值比较：进行方差齐性检验,得到P=0.629,不拒绝H0,方差齐。方差分析结果：整体比较5组间差异存在统计学意义(P=0.000),需要进一步做多重比较。结果显示Scheffe法和Bonferroni法结果一致,即21-30岁年龄组与41-50岁年龄组(P=0.000)、21-30岁年龄组和51-60岁年龄组(P=0.000)、21-30岁年龄组和61-70岁年龄组(P=0.000)、31-40岁年龄组和41-50岁年龄组(P值分别为0.002和0.001)、31-40岁年龄组和51-60岁年龄组(P=0.000)、31-40岁年龄组和61-70岁年龄组(P=0.000)之间差异有统计学意义。LSD法结果显示除了上述组别的差异具有统计学意义以外,41-50岁年龄组和61-70岁年龄组差异具有统计学意义(P=0.007)。而SNK法结果显示除去21-30岁年龄组与31-40岁年龄组(P=0.070)、41-50岁年龄组与51-60岁年龄组(P=0.099)以及51-60岁年龄组与61-70岁年龄组(P=0.290)差异无统计学意义以外,其它组之间差异均有统计学意义。通过对比可知LSD法和SNK法的结果一致。男性各组CT值最小值比较：进行方差齐性检验,得到P=0.000,拒绝H0,方差不齐。方差分析结果：多重比较选择基于方差不齐的多重比较方法Dunnett'sT3和Dunnett's C。Dunnett's T3法：21-30岁年龄组与51-60岁年龄组(P=0.003)、21-30岁年龄组与61-70岁年龄组(P=0.000)、41-50岁年龄组与61-70岁年龄组(P=0.003)以及51-60岁年龄组与61-70岁年龄组(P=0.049)差异存在统计学意义,其它各组均无统计学意义。Dunnett's C法：该方法只给出区间估计结果。对于可信区间的结果,若可信区间不包含0则表示有统计学差异,21-30岁年龄组与41-50岁年龄组(0.5317～104.8283)、21-30岁年龄组与51-60岁年龄组(26.4708～167.3692)、21-30岁年龄组与61-70岁年龄组(115.2617～400.0583)、41-50岁年龄组与61-70岁年龄组(54.9861～354.9739)以及51-60岁年龄组与61-70岁年龄组(3.4447-318.0353)差异存在统计学意义,其它各组均无统计学意义；男性各组CT值最大值比较：进行方差齐性检验,得到P=0.158,不拒绝H0,方差齐。方差分析结果：整体比较5组间不存在统计差异(P=0.196)。女性所有CT值指标总体方差均不齐,但整体组间差异具有统计学意义,需要进一步作多重比较。女性各组CT值平均值比较：Dunnett's T3法：21-30岁年龄组与51-60岁年龄组(P=0.000)、21-30岁年龄组与61-70岁年龄组(P=0.000)、31-40岁年龄组与51-60岁年龄组(P=0.000)、31-40岁年龄组与61-70岁年龄组(P=0.000)、41-50岁年龄组与51-60岁年龄组(P=0.002)、41-50岁年龄组与61-70岁年龄组(P=0.000)以及51-60岁年龄组与61-70岁年龄组(P=0.000)之间差异存在统计学意义,其它各组均无统计学意义。Dunnett's C法：对于可信区间的结果,若可信区间不包含0则表示有统计学差异,21-30岁年龄组与51-60岁年龄组(33.4098-108.2318)、21-30岁年龄组与61-70岁年龄组(104.8592～164.9632)、31-40岁年龄组与51-60岁年龄组(38.0154-119.9654)、31-40岁年龄组和61-70岁年龄组(108.6934～177.4682)、41-50岁年龄组和51-60岁年龄组(14.1076-90.6100)、41-50岁年龄组和61-70岁年龄组(85.3575～147.5409)、51-60岁年龄组和61-70岁年龄组(25.7676～102.4132)之间差异存在统计学意义,其它各组均无统计学意义。女性各组CT值最小值比较：Dunnett's T3法：21-30岁年龄组与41-50岁年龄组(P=0.034)、21-30岁年龄组与51-60岁年龄组(P=0.001)、21-30岁年龄组与61-70岁年龄组(P=0.000)、31-40岁年龄组与51-60岁年龄组(P=0.016)、31-40岁年龄组与61-70岁年龄组(P=0.002)、41-50岁年龄组与51-60岁年龄组(P=0.017)以及41-50岁年龄组与61-70岁年龄组(P=0.001)之间差异存在统计学意义,其它各组均无统计学意义。Dunnett's C法：对于可信区间的结果,若可信区间不包含0则表示有统计学差异,21-30岁年龄组与41-50岁年龄组(1.6356-53.2044)、21-30岁年龄组与51-60岁年龄组(43.0432～213.3568)、21-30岁年龄组与61-70岁年龄组(61.0564～197.1836)、31-40岁年龄组与51-60岁年龄组(15.3406～218.2594)、31-40岁与61-70岁年龄组(30.1131～205.3269)、41-50岁年龄组和51-60岁年龄组(13.8947～187.6653)、41-50岁年龄组和61-70岁年龄组(31.4859-171.9141)之间差异存在统计学意义,其它各组均无统计学意义。女性各组CT值最大值比较：Dunnett's T3法：21-30岁年龄组与61-70岁年龄组(P=0.003)、31-40岁年龄组与61-70岁年龄组(P=0.000)、41-50岁年龄组与61-70岁年龄组(P=0.000)以及51-60岁年龄组与61-70岁年龄组(P=0.017)之间差异存在统计学意义,其它各组均无统计学意义。Dunnett's C法：对于可信区间的结果,若可信区间不包含0则表示有统计学差异,21-30岁年龄组与61-70岁年龄组(9.1578～66.5622)、31-40岁年龄组与61-70岁年龄组(28.2892～83.5508)、41-50岁年龄组与61-70岁年龄组(26.3460～83.6540)、51-60岁年龄组与61-70岁年龄组(4.4480～67.9120)之间差异存在统计学意义,其它各组均无统计学意义。当读取的男性和女性不同年龄段组的CT值采用平均值±标准差表示时,对相同年龄段男性和女性腰椎CT值平均值进行比较,结果显示：除去21-30岁年龄组以及51-60岁年龄组以外,其它组间差异均具有统计学意义。分别对男性以及女性不同年龄段腰椎CT值平均值进行比较,结果显示：男性组：21-30岁组、31-40岁组分别与41岁以上各年龄组以及41-50岁组与61-70岁组比较差异有统计学意义；女性组：21-30岁组、31-40岁组、41-50岁组分别与51岁以上各年龄组以及51-60岁组与61-70岁组比较差异有统计学意义。结论 1.CT值平均值与骨密度值具有相关性。 2.通过新的图像分割方法可以快速准确获取任意感兴趣区域的腰椎CT值。 3.腰椎CT值平均值随着年龄增加而降低。
[Abstract]:background
In clinical work, a large number of patients with lumbago and leg pain are found, some are even very young and are often seen in men. But the clinicians often take some routine examinations for the physical examination of these patients. In the case of certain positive signs, the patients are given X-ray or CT, even MRI, and other imaging examinations, It is often the diagnosis of lumbar disc herniation, protrusion, etc. until the patient has a typical symptom and even serious complications such as fracture, the exact diagnosis of osteoporosis is obtained. This time not only increases the burden of life, reduces the quality of life, but also increases the financial burden of the patient. A new image segmentation method is used to separate each vertebra CT image from each vertebra one by one and read the maximum value, minimum and average of the CT value of the vertebral body model. The minimum value of the CT value of a part of the vertebral body is negative. At the lowest point, we can even reach -1024. that we speculate that these vertebral bodies should contain There is a gas component, and try to find the cause of the gas in the vertebral body. In general, the elderly patients may have different degrees of lumbar degeneration, there may be gas in the vertebral body, but there is no degenerative disease in the young patients. Therefore, we speculate that there may be some other lesions in the young patients' lumbar vertebrae. The step survey found that the patients' relatives had a history of primary osteoporosis, and we speculated that the patient may have some symptoms of primary osteoporosis such as reduced bone density.
objective
The lumbar spiral CT scan image of lumbar and leg pain of men and women of different ages is introduced into the 3D reconstruction software. The lumbar vertebra structure in these images is divided by a new image segmentation method. The maximum value, minimum value and average value of lumbar CT value are solved and the change law is explored, and the CT value and bone density are explored. To achieve the purpose of screening young people, early diagnosis of those with untypical symptoms, early treatment, reducing the patient's financial burden and living burden, reducing the rate of disability and complications, and improving the quality of life.
Method
202 cases of lumbar and leg pain at the age of 21-70 years were randomly collected and divided into groups according to age at 10 years of age. Each age group was divided into two groups of men and women, including 15 men and women in 21-30 years of age, 22 in 31-40 years of age, 21 in 41-50 years of age, and 21 for men and women of 51-60 age groups. 22 cases of men and women at the age of 61-70, each group of male and female 50 vertebral bodies, with a total of 500 vertebrae. The patients were scanned for spiral CT. The CT device was 120KV, 264MAS, and 0.625mm in the 64 row CT scanner of the American Ge Corp. The image was recorded in DICOM format of 512x512 pixels to the output of DVD disc. First, a new map was used. A total of 500 vertebral bodies, 202 cases of lumbar and leg pain 21-70 years old, were separated by three-dimensional reconstruction software, and the maximum value, minimum and average value of the CT value of each age group of male and female patients were calculated, and the statistical analysis was performed respectively.
Under the Windows7Ultimate64bit operating system, the CT scanning data was introduced into the medical 3D reconstruction software Mimics14.0. and then the 3D model of the lumbar spine was reconstructed. The ilium and the sacrum were removed first. Then the vertebral bodies were separated and the corresponding independent mask (Mask) files were separated and separated from the vertebral body, the pedicle and the vertebral lamina, and the closed space around the vertebral body. Then the model contour is calculated, the holes in the Mask are filled and the lumbar vertebrae of each patient are divided by the Boolean operation. The average value, maximum and minimum of the CT value of the model of the vertebral body are completed.
The experimental data were statistically processed with SPSS16.0 software, and a single factor variance analysis was used to compare the maximum, minimum and average value of the CT value of the men and women with low back and leg pain in each age group. When the CT value of the men and women in different age groups was expressed with the average standard deviation of the standard deviation, we were also compared. The mean value of the CT value of the male and female lumbar vertebrae in the same age group and the mean value of the CT value of the male and female lumbar vertebrae in different age groups were examined by single factor analysis of variance for the maximum, minimum and average value of the CT value of the men and women of different age groups, and the difference was statistically significant. If there was a statistical difference, the difference was statistically significant. The multiple comparison between groups was carried out. When the multiple comparison between groups was carried out, the variance homogeneity test was performed on all parts of the data. LSD, Bonferroni, SNK and Scheffe were used when the variance was homogeneous. The Dunnett's T3 method and Dunnett's C method were used when the variance was inhomogeneous, and P0.05 believed that the difference had the significance of unified planning.
Result
The average value of CT in male groups was compared: P=0.629, H0, Fang Chaqi. Analysis of variance: the difference between the 5 groups was statistically significant (P=0.000), and multiple comparisons were needed. The results showed that the results of Scheffe and Bonferroni were the same, that is, the age group of 21-30 years and the age group of 41-50 years (P=0). .000), 21-30 year old age group and 51-60 year age group (P=0.000), 21-30 year old age group and 61-70 year age group (P=0.000), 31-40 year old age group and 41-50 year old group (P value 0.002 and 0.001 respectively), 31-40 year old age group and 51-60 year old age group (P=0.000), the difference between 31-40 age group and 61-70 age group (P=0.000) is statistically significant.LSD knot The results showed that the differences of the 41-50 year old age group and the 61-70 year age group were statistically significant (P=0.007). The results of the SNK method showed that the 21-30 year old age group and the 31-40 year old age group (P=0.070), the 41-50 year old group and the 51-60 year old group (P=0.099), the 51-60 year old age group and the 61-70 year old age group were found. There was no significant difference between the two groups (P=0.290). The difference between the other groups was statistically significant. By contrast, the results of LSD and SNK were consistent.
The minimum value of CT values in male groups: a homogeneity test of variance, P=0.000, rejection of H0, variance of variance. Variance analysis results: multiple comparison selection based on multiple comparison method based on variance inhomogeneous Dunnett'sT3 and Dunnett's C.Dunnett's T3 method: 21-30 year old age group and 51-60 year old age group (P=0.003), 21-30 year old age group and 61-70 year old age Group (P=0.000), 41-50 year old age group and 61-70 year age group (P=0.003), 51-60 year old age group and 61-70 year age group (P=0.049) have statistical significance, other groups have no statistical significance.Dunnett's C method: this method only gives interval estimation results. For the result of the confidence interval, if the confidence interval does not contain 0, there are statistics. The 21-30 year old age group and 41-50 year old age group (0.5317 to 104.8283), 21-30 year old age group and 51-60 age group (26.4708 to 167.3692), 21-30 year old age group and 61-70 year old age group (115.2617 ~ 400.0583), 41-50 years age group and age group, age group and age group (3.4447-318.035 3) the difference was statistically significant, but no significant difference was found in the other groups.
The maximum value of CT values in male groups was compared: the homogeneity test of variance, P=0.158, no refusal of H0, Fang Chaqi. Results of variance analysis: there was no statistical difference between the 5 groups (P=0.196).
The overall variance of all CT values of females was not uniform, but the difference between the overall groups was statistically significant. Further multiple comparisons were needed.
The average value of CT values in female groups: Dunnett's T3: 21-30 year old age group and 51-60 year age group (P=0.000), 21-30 year old age group and 61-70 year age group (P=0.000), 31-40 year old age group and 51-60 year age group (P=0.000), 31-40 year old age group and 61-70 year age group (P= 0), 41-50 years age group and 51-60 year age group (P=0.002), 41-50 year old. There was significant difference between the age group and the age group of 61-70 years (P=0.000), 51-60 year old age group and 61-70 year old age group (P=0.000), and there was no statistical significance.Dunnett's C method in the other groups. For the results of the confidence interval, if the confidence interval did not contain 0, there were statistical differences, 21-30 year old age group and 51-60 year old group (33.4098- 108.2318) 21-30 years old age group and 61-70 year old age group (104.8592 ~ 164.9632), 31-40 year old age group and 51-60 year age group (38.0154-119.9654), 31-40 age group and 61-70 year old age group (108.6934 ~ 177.4682), 41-50 age group and 51-60 age group (14.1076-90.6100), age group and age group. The difference was statistically significant between the -60 age group and the 61-70 year old group (25.7676 to 102.4132), but no significant difference was found in the other groups.
The minimum value comparison of CT values in female groups: Dunnett's T3: 21-30 year old age group and 41-50 year age group (P=0.034), 21-30 year old age group and 51-60 year age group (P=0.001), 21-30 year old age group and 61-70 year age group (P=0.000), 31-40 year old age group and 51-60 year age group (P= 0.016), 31-40 years age group and 61-70 year age group (P=0.002), 41-50 year old. There was significant difference between the age group and the age group of 51-60 years (P=0.017), 41-50 year old age group and 61-70 year old age group (P=0.001), and there was no statistical significance.Dunnett's C method in the other groups. For the results of the confidence interval, if the confidence interval did not contain 0, there were statistical differences, 21-30 year old age group and 41-50 year old group (1.6356-5 3.2044) 21-30 years old age group and 51-60 year old age group (43.0432 ~ 213.3568), 21-30 year old age group and 61-70 year age group (61.0564 ~ 197.1836), 31-40 years old age group and 51-60 age group (15.3406 ~ 218.2594). The difference between age group and 61-70 year old age group (31.4859-171.9141) was statistically significant, and no significant difference was found in other groups.
The maximum value comparison of CT values in female groups: Dunnett's T3 method: 21-30 year old age group and 61-70 year age group (P=0.003), 31-40 year old age group and 61-70 year age group (P=0.000), 41-50 year old age group and 61-70 age group (P=0.000), 51-60 year old age group and 61-70 age group (P=0.017), and the other groups have no significant difference. Statistical significance.Dunnett's C method: for the results of the confidence interval, if the confidence interval does not contain 0 statistical differences, 21-30 year old age group and 61-70 year old age group (9.1578 ~ 66.5622), 31-40 age group and 61-70 age group (28.2892 to 83.5508), 41-50 years old and 61-70 years of age group (26.3460 ~ 83.6540), the age of 51-60 There was statistically significant difference between the group and the 61-70 year old group (4.4480 to 67.9120), but no significant difference was found in the other groups.
When the CT values of the men and women in different age groups were represented by average mean standard deviation, the mean values of the CT values of the male and female lumbar vertebrae in the same age group were compared. The results showed that the difference between the 21-30 years old age group and the 51-60 age group was statistically significant.
The average values of lumbar CT values in male and female different age groups were compared. The results showed that: male group: 21-30 year old group, 31-40 year old group compared with 41 years old group, 41-50 year old group and 61-70 year old group, respectively. The female group: 21-30 year old group, 31-40 year old group, 41-50 year old group and all ages above 51 years old, respectively. The difference between the group and the 51-60 year old group was statistically significant compared with that of the 61-70 year old group.
conclusion
The mean value of 1.CT value is correlated with the bone mineral density.
2. through the new image segmentation method, the lumbar CT value of any region of interest can be obtained quickly and accurately.
The average CT value of the lumbar vertebra decreased with the increase of age.
【学位授予单位】：南方医科大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：R681.5;R816.8

【参考文献】