CT Grade | Cases | Ground-glass opacity | Lobular septal thickening | Grid-like shadow | Consolidation shadow | Honeycomb shadow |
Mild | 13 | 69.2% | 84.6% | 23.1% | 0 | 0 |
Moderate | 31 | 64.5% | 100% | 93.5% | 25.8% | 6% |
Severe | 16 | 93.8% | 100% | 100% | 50% | 87.5% |
Citation: | ZHANG Z, LI M, ZHAO L Q, et al. Evaluation on the hepatic functional reserve of cirrhotic patients based on CT radiomics characteristics[J]. CT Theory and Applications, 2022, 31(1): 55-62. DOI: 10.15953/j.ctta.2021.019. (in Chinese). |
Objective: To explore the value of CT radiomics characteristics in evaluating hepatic functional reserve of patients with liver cirrhosis. Methods: A total of 121 patients with clinically confirmed liver cirrhosis were retrospectively collected, who were graded as A, B and C according to Child-Pugh standard. 51 cases were grade A (group A) and 70 cases were grade B and C (group B). All patients underwent non-contrast and contrast enhanced CT scan using GE Discovery CT 750 HD scanner. Enhanced CT images of portal venous phase were selected, and the whole liver was delineated at the level of the left portal vein by 2 radiologists using Shukun Radiomics V94 software. Intraclass correlation coefficient (ICC) was applied to test the inter-group consistency of the results obtained by the 2 radiologists. All patients were randomly divided into the training set and the validation set at a ratio of 7∶3, and the whole liver was performed extraction process for radiomics characteristics. After dimensionality reduction, radiomics characteristics were acquired and the model was established by Logistic Regression (LR). Area Under ROC Curve (AUC) was used to evaluate the performance of the model. Results: The consistency test results of image delineation by two radiologists turned out well and the ICC were greater than 0.75. Finally, 17 radiomics characteristics were used to establish the evaluation model for hepatic functional reserve of patients with liver cirrhosis. The AUC of the training set was 0.84 while the accuracy, sensitivity and specificity was 0.78, 0.79, and 0.77, respectively. The AUC of the validation set was 0.77 while the accuracy, sensitivity and specificity was 0.71, 0.76, and 0.65, respectively. Conclusion: CT radiomics characteristics could be used to evaluate the hepatic functional reserve of patients with liver cirrhosis.
Interstitial lung disease (ILD) is a kind of disease characterized by interstitial inflammation and fibrosis. There are many causes[1-2] for ILD, a common disease that seriously endangers human health, which mainly involves the interstitium and alveolar cavity, leading to the loss of alveolar-capillary functional units[3]. Pulmonary function and medical imaging examination play an important role in the diagnosis and treatment of ILD, CT examination is simple and feasible with objective images, which can provide a qualitative diagnosis of interstitial lung disease[4], but it cannot further understand the patient’s lung function. Pulmonary function tests are of great significance for determining airflow limitation, judging the condition and prognosis.
However, pulmonary function tests are easily affected by subjective and other factors, and for some patients with severely impaired lung function, it is often impossible to complete routine pulmonary function tests. However, CT severity classification of ILD is rarely reported either nationally or internationally. Here we aim to grade the CT severity of the disease by combining the pulmonary function with CT examination through CT scoring, therefore creating a new basis for the clinical evaluation of the disease and providing help for the clinical diagnosis and treatment.
In this retrospective study, we evaluated 60 patients with clinical diagnosis of ILD from September 2018 to January 2020 in our hospital. After taking an informed consent from each patient, detailed demographic data, clinical history and examination, and laboratory data were reviewed, along with the CT findings.All patients had undergone chest CT and pulmonary function examination, and the following exclusion criteria were used:
(1) Unable to take a deep breath and hold it; (2) Previous history of chest surgery or restrictive lung diseases (such as granuloma, tumor, tuberculosis); (3) Serious heart disease and pulmonary hypertension; (4) Acute interstitial pneumonia.
The final study group consisted of 60 patients (36 males and 24 females, mean age (63 ±14) years; age range 27~85 years. The main clinical symptoms are cough, expectoration, shortness of breath, chest tightness, chest pain.
In all 60 patients, HRCT scan was performed on a 128-slice spiral CT scanner (Siemens). Prior to CT scanning each patient was subjected to breathing training and was told to hold his/her breath after deep inhalation. The patient was then scanned from the tip of the lung to the bottom of the lung. Scanning parameters used were as follows:
Voltage: 120 kV, tube current: intelligent tube current, screw pitch:1.5, acquisition matrix: 512×512, canning acquisition layer thickness: 0.75 mm, reconstruction layer thickness: 1 mm, layer spacing: 1 mm.
Two senior radiologists who were blinded to the patient outcome parameters, reviewed all imaging studies and recorded all findings. They independently scored the severity grading of all patients, and any differences between the two readers were subsequently resolved by consensus to obtain a consensus score. Grading standard was based on Camiciottoli4, which composed of pathological changes of lung parenchymal lesion and their respective ranges: the former include ground-glass opacity (score 1), the shadow of consolidation of the lung (score 2), irregular increasing and thickening of the interlobular septa (score 3), grid-like shadow (score 4), honeycomb shadow (score 5), lesion type score from 0 (no abnormalities) to 15 (all 5 abnormalities were detected); According to the distribution of bronchial lung segment left and right lung lobes can be divided into 10 pulmonary segments each, resulting in a total of 20 lung segments.
The score is 1 point for 1~3 affected lung segments, 2 points for 4~9 lung segment, 3 points for more than 9 affected lung segments, lesion involvement range scores from 0 point (no abnormalities in lung segments) to 15 points (all 5 lesion types are involved which affected more than 9 lung segments). The two scores are combined to obtain a total score (0~30 points). At the same time, the author innovatively proposed to grade the severity of patients according to CT scores. Patients with a score of 0 to 10 were defined as CT severity grades as mild, 11 to 20 as moderate, and 21 to 30 as severe.
Pulmonary function test (PFT) was performed with the lung function detector of JAEGER, Germany. The patient was seated during the examination, and the one-second forced breathing volume and CO dispersion were measured.
According to the patient’s body mass,age, sex and height, the predicted value was obtained, and the measured value/predicted value of the above PFT was obtained (real/expected) (%). The patient’s PFT was taken before and within three days after the CT scan.
The data were expressed as (mean ± standard) deviation and statistically analyzed using SPSS 19.0 software.
Pearson correlation test requires variables to be normally distributed, and was used to determine the correlation between CT score of each patient and FEV1% and DLCO%; The difference of FEV1 and DLCO values in patients with different CT severity classification was analyzed by one-way ANOVA, and the sample mean of the three values was analyzed by pairwise comparison of LSD-t test. Double-layer test was used for all statistical analysis, and P < 0.05 was considered statistically significant.
Frequency of various signs in 60 patients is summarized as follows: In 13 patients with mild symptoms, the ground-glass opacity was 69.2% (9/13), lobular septal thickening was 84.6% (11/13), grid-like shadow was 23.1% (3/13), and consolidation shadow and honeycomb shadow were not detected. Among the 31 patients graded as moderate, the ground-glass opacity was 64.5% (20/31), the lobular septal thickening was 100% (31/31), the grid-like shadow was 93.5% (29/31), the consolidation shadow was 25.8% (8/31), and the honeycomb shadow was 6% (2/31). Among the 16 patients with severe grade, the ground-glass opacity was 93.8% (15/16), the lobular septal thickening was 100% (16/16), the grid-like shadow was 100% (16/16), the consolidation shadow was 50% (8/16), and the honeycomb shadow was 87.5% (14/16) (Table 1).
CT Grade | Cases | Ground-glass opacity | Lobular septal thickening | Grid-like shadow | Consolidation shadow | Honeycomb shadow |
Mild | 13 | 69.2% | 84.6% | 23.1% | 0 | 0 |
Moderate | 31 | 64.5% | 100% | 93.5% | 25.8% | 6% |
Severe | 16 | 93.8% | 100% | 100% | 50% | 87.5% |
Among the 60 patients, there were 13 patients with 1~10 points, 31 patients with 11~20 points, and 16 patients with 21~30 points.
CT score was negatively correlated with DLCO% and FEV1%, with r being −0.814 and −0.797, respectively.
As mentioned earlier, 13 patients with CT scores of 1~10 was defined as mild, 31 patients with CT scores of 11~20 was defined as moderate, and 16 patients with 21~30 CT scores were categorized as severe cases (Fig.1). The mean values of DLCO% and FEV1% for 13 mild patients were (67.61±6.46) and (87.99±6.48), 31 moderate patients were (59.97±8.13) and (78.38±10.19), and 16 severe patients were (35.60±15.03) and (62.25±8.08), respectively. The pairwise comparison of the mean values between groups was statistically significant (Table 2).
CT Grade DLCO% | n | Cases FEV1% | |
Mild | 13 | 67.61±6.46 | 87.99±6.48 |
Moderate | 31 | 59.97±8.13 | 78.38±10.19 |
Severe | 16 | 35.60±15.03 | 62.25±8.08 |
The pulmonary function parameters of mild, moderate and severe groups were pairwise compared with each other (P < 0.01). |
ILD is a common clinical disease with numerous etiologies. Chest CT and pulmonary function test play an important role in the diagnosis, treatment and evaluation of the outcome of the disease[5-6]. According to the involved sites, imaging included interstitial and substantial lesions, the former mainly involved the interstitial alveolar wall, interlobular septum, interlobular septum, connective tissue around the bronchovascular bundle and pleura, interlobular fissure and other structures, the latter mainly involved the widely adjacent alveolar cavity. The corresponding CT signs are mainly reflected as ground-glass opacity, lobular septal thickening, grid-like shadow, consolidation shadow and honeycomb shadow. Previous studies have shown that ground-glass opacity and interlobular septal thickening are the early manifestations of ILD, while grid-like shadow and honeycomb shadow often indicate that the disease has developed to the middle, late and even irreversible stages[7-8]. In this group of cases, the occurrence frequency of ground-glass opacity and interlobular septum in mild patients were 69.2% and 84.6%, the occurrence frequency of grid-like shadow in moderate patients were 93.5%, and the occurrence of honeycomb shadow in severe patients were 87.5%. From another aspect, it was confirmed that the CT severity grading of ILD patients in this study was objective, accurate and consistent with clinical practice.
In addition, in this study, patients with mild ILD showed no consolidation shadows, and the occurrence frequency of such in moderate and severe patients was 25.8% and 50%. It showed that lung consolidation may also indicate the severity of patients to a certain extent; lung consolidation was a direct sign of acute inflammation, and it also suggested that moderate and severe patients were more likely to have acute infections than mild patients. Pulmonary function of a typical ILD is characterized by decreased diffusion function, restrictive ventilation disorder, and decreased lung volume[9-10].
In this study, CT scores were obtained by different imaging signs and distribution of ILD, and correlation analysis was conducted between CT scores and FEV1% and DLCO% indexes of pulmonary function, suggesting that CT scores were negatively correlated with both, with correlation coefficients of −0.814 and −0.797, respectively, suggesting a good correlation. FEV1% was the main index to judge the restricted ventilation disorder[11-12], and the CT score was negatively correlated with it, showing a good correlation, indicating that the value of FEVI decreased with the increase of CT score, and that the CT score could reflect the restricted ventilation disorder of ILD. In addition, reduced diffusion function is a relatively sensitive diagnostic index for early changes in ILD, while the ILD of carbon monoxide is mainly manifested as diffuse dysfunction, and the CT score is negatively correlated with DLCO, indicating that the CT score can be used to evaluate the lung function of patients to a certain extent.
On the basis of the CT score, the author defined the CT severity of patients from 0 to 10 as mild, from 11 to 20 as moderate, and from 21 to 30 as severe, the difference between FEV1% and DLCO% value in the three groups was statistically significant, which indicated that there was a good basis for the severity grading of ILD patients by CT score, which was consistent with the judgment of lung function index on the disease. In this way, in clinical work, CT severity grading can provide a simpler and faster method for clinical and radiologist to judge disease severity. However, in this study, only 60 patients have complete FEV1%, DLCO% data as lung function index, TLC, VC, RV that reflect changes in lung volume were not included in the evaluation and comparison of CT severity classification. It is unclear if CT severity grading can accurately reflect changes in lung volume. Many researchers have included quantitative analysis of CT pulmonary function test in the evaluation of ILD[13-16]. We will further analyze the CT severity classification and CT pulmonary function indicators, and at the same time, more patient data and conventional pulmonary function indicators should be added to classify the CT severity of pulmonary interstitial disease for increased objectivity and accuracy.
The authors declare no conflict of interest.
Funding was providing by the Shenzhen City Nanshan District Science and Technology Plan Project (Medical and Health Category) (Grant No.2018073).
[1] |
田秋菊, 贾继东. 肝硬化腹水及其相关并发症治疗新进展[J]. 实用肝脏病杂志, 2015,18(1): 1−3. doi: 10.3969/j.issn.1672-5069.2015.01.001
TIAN Q J, JIA J D. New progress in the treatment of cirrhotic ascites and related complications[J]. Journal of Practical Liver Diseases, 2015, 18(1): 1−3. (in Chinese). doi: 10.3969/j.issn.1672-5069.2015.01.001
|
[2] |
GWAK G Y. Renewed 2015 clinical practice guidelines for management of hepatitis C by Korean association for the study of the liver; What has been changed? Treatment of patients with decompensated cirrhosis[J]. The Korean Journal of Gastroenterology = Taehan Sohwagi Hakhoechi, 2016, 67(3): 137−141. doi: 10.4166/kjg.2016.67.3.137
|
[3] |
周志华. 机器学习[M]. 北京: 清华大学出版社, 2016.
ZHOU Z H. Machine learning[M]. Beijing: Tsinghua University Press, 2016. (in Chinese).
|
[4] |
王宇, 王民, 张冠华, 等. 肝硬化的诊断, 分期及治疗原则[J]. 临床肝胆病杂志, 2021,37(1): 5.
WANG Y, WANG M, ZHANG G H, et al. Diagnosis, staging and treatment principles of liver cirrhosis[J]. Journal of Clinical Hepatobiliary Diseases, 2021, 37(1): 5. (in Chinese).
|
[5] |
杨乔怡, 王恩鑫, 王哲轩, 等. 经肝动脉化疗栓塞术治疗伴门静脉癌栓肝细胞癌的效果及预后评价模型分析[J]. 临床肝胆病杂志, 2021,37(3): 611-615.
YANG Q Y, WANG E X, WANG Z X, et al. Clinical effect of transcatheter arterial chemoembolization in treatment of patients with hepatocellular carcinoma and portal vein tumor thrombus and an analysis of prognosic evaluation models[J]. Journal of Clinical Hepatobiliary Diseases, 2021, 37(3): 611-615. (in Chinese).
|
[6] |
PANTEA R, MEISTER P, NEUHAUS J P, et al. Chirurgie bei patienten mit leberzirrhose surgery in patients with liver cirrhosis[J]. Chirurg, 2021, 92(9): 838−845. doi: 10.1007/s00104-020-01319-z
|
[7] |
张豪, 邹立秋. 磁共振弹性成像技术对肝纤维化诊断的新进展[J]. 磁共振成像, 2020,94(4): 85−87.
ZHANG H, ZOU L Q. New progress of magnetic resonance elastography in the diagnosis of liver fibrosis[J]. Magnetic Resonance Imaging, 2020, 94(4): 85−87. (in Chinese).
|
[8] |
SINGH R, WILSON M P, KATLARIWALA P, et al. Accuracy of liver and spleen stiffness on magnetic resonance elastography for detecting portal hypertension: A systematic review and meta-analysis[J]. European Journal of Gastroenterology & Hepatology, 2021, 32(2): 237−245.
|
[9] |
STANKOVIC Z. Four-dimensional flow magnetic resonance imaging in cirrhosis[J]. World Journal Gastroenterology, 2016, 22(1): 89−102. doi: 10.3748/wjg.v22.i1.89
|
[10] |
HAN X, AN W, CAO Q, et al. Noninvasive evaluation of esophageal varices in cirrhotic patients based on spleen hemodynamics: A dual-energy CT study[J]. European Radiology, 2020, 30(6): 3210−3216. doi: 10.1007/s00330-020-06680-5
|
[11] |
刘林祥, 聂源, 朱萱. 瞬时弹性成像技术在肝硬化及其并发症中的应用[J]. 临床肝胆病杂志, 2020, 36(6): 172-175.
LIU L X, NIE Y, ZHU X. Application of transient elastography in liver cirrhosis and its complications[J]. Journal of Clinical Hepatobiliary Diseases, 2020, 36(6): 172-175. (in Chinese).
|
[12] |
ATTENBERGER U I, LANGS G. How does radiomics actually work? -Review[J]. Rofo, 2021, 193(6): 652−657. doi: 10.1055/a-1293-8953
|
[13] |
RAMAN S P, SCHROEDER J L, HUANG P, et al. Preliminary data using computed tomography texture analysis for the classification of hypervascular liver lesions: Generation of a predictive model on the basis of quantitative spatial frequency measurements-a work in progress[J]. Journal of Computer Assisted Tomography, 2015, 39(3): 383−395.
|
[14] |
DAVNALL F, YIP C, LJUNGQVIST G, et al. Assessment of tumor heterogeneity: An emerging imaging tool for clinical practice?[J]. Insights Into Imaging, 2012, 3(6): 573−589. doi: 10.1007/s13244-012-0196-6
|
[15] |
BALAJI G, MILES K A. Quantifying tumour heterogeneity with CT[J]. Cancer Imaging the Official Publication of the International Cancer Imaging Society, 2013, 13(1): 140−149.
|
[16] |
HUSSAIN M A, HAMARNEH G, GARBI R. Learnable image histograms-based deep radiomics for renal cell carcinoma grading and staging[J]. Computerized Medical Imaging and Graphics, 2021, 90: 101924. doi: 10.1016/j.compmedimag.2021.101924
|
[17] |
LUBNER M G, SMITH A D, SANDRASEGARAN K, et al. CT texture analysis: Definitions, applications, biologic correlates, and challenges[J]. RadioGraphics, 2017, 37(5): 1483−1503. doi: 10.1148/rg.2017170056
|
[18] |
中华医学会肝病学分会. 肝硬化诊治指南[J]. 中华肝脏病杂志, 2019,27(11): 846−865. doi: 10.3760/cma.j.issn.1007-3418.2019.11.008
Hepatology Branch of Chinese Medical Association. Guidelines for the diagnosis and treatment of liver cirrhosis[J]. Chinese Journal of Liver Diseases, 2019, 27(11): 846−865. doi: 10.3760/cma.j.issn.1007-3418.2019.11.008
|
[19] |
LUBNER M G, MALECKI K, KLOKE J, et al. Texture analysis of the liver at MDCT for assessing hepatic fibrosis[J]. Abdominal Radiology, 2017, 42(8): 2069−2078. doi: 10.1007/s00261-017-1096-5
|
[20] |
TAKESHI Y, TANYA W, KEIKO I, et al. Evaluation of liver fibrosis using texture analysis on combined-contrast-enhanced magnetic resonance images at 3.0T[J]. Biomed Research International, 2015, 2015(5): 1−12.
|
[21] |
ZHANG X, GAO X, LIU B J, et al. Effective staging of fibrosis by the selected texture features of liver: Which one is better, CT or MR imaging?[J]. Computerized Medical Imaging and Graphics, 2015, 46(2): 227−236.
|
[22] |
WANG J C, FU R, TAO X W, et al. A radiomics-based model on non-contrast CT for predicting cirrhosis: make the most of image data[J]. Biomarker Research, 2020, 8(1): 47. doi: 10.1186/s40364-020-00219-y
|
1. |
卢振如,王婷. 成人肠道套细胞淋巴瘤合并回盲型肠套叠一例. 中国医学科学院学报. 2024(03): 458-461 .
![]() | |
2. |
王婷,田松林,郑珂. MR小肠造影对肠道T细胞淋巴瘤与B细胞淋巴瘤的鉴别效能分析. 罕少疾病杂志. 2024(10): 95-97 .
![]() | |
3. |
陈衍池. 超声诊断肠梗阻的价值分析. 中国医疗器械信息. 2023(14): 16-18+176 .
![]() |
CT Grade | Cases | Ground-glass opacity | Lobular septal thickening | Grid-like shadow | Consolidation shadow | Honeycomb shadow |
Mild | 13 | 69.2% | 84.6% | 23.1% | 0 | 0 |
Moderate | 31 | 64.5% | 100% | 93.5% | 25.8% | 6% |
Severe | 16 | 93.8% | 100% | 100% | 50% | 87.5% |
CT Grade DLCO% | n | Cases FEV1% | |
Mild | 13 | 67.61±6.46 | 87.99±6.48 |
Moderate | 31 | 59.97±8.13 | 78.38±10.19 |
Severe | 16 | 35.60±15.03 | 62.25±8.08 |
The pulmonary function parameters of mild, moderate and severe groups were pairwise compared with each other (P < 0.01). |