Chin J Cancer Res

Deep learning model improves radiologists’ performance in detection and classification of breast lesions

doi: 10.21147/j.issn.1000-9604.2021.06.05

Objective Computer-aided diagnosis using deep learning algorithms has been initially applied in the field of mammography, but there is no large-scale clinical application. Methods This study proposed to develop and verify an artificial intelligence model based on mammography. Firstly, mammograms retrospectively collected from six centers were randomized to a training dataset and a validation dataset for establishing the model. Secondly, the model was tested by comparing 12 radiologists’ performance with and without it. Finally, prospectively enrolled women with mammograms from six centers were diagnosed by radiologists with the model. The detection and diagnostic capabilities were evaluated using the free-response receiver operating characteristic (FROC) curve and ROC curve. Results The sensitivity of model for detecting lesions after matching was 0.908 for false positive rate of 0.25 in unilateral images. The area under ROC curve (AUC) to distinguish the benign lesions from malignant lesions was 0.855 [95% confidence interval (95% CI): 0.830, 0.880]. The performance of 12 radiologists with the model was higher than that of radiologists alone (AUC: 0.852 vs. 0.805, P=0.005). The mean reading time of with the model was shorter than that of reading alone (80.18 s vs. 62.28 s, P=0.032). In prospective application, the sensitivity of detection reached 0.887 at false positive rate of 0.25; the AUC of radiologists with the model was 0.983 (95% CI: 0.978, 0.988), with sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 94.36%, 98.07%, 87.76%, and 99.09%, respectively. Conclusions The artificial intelligence model exhibits high accuracy for detecting and diagnosing breast lesions, improves diagnostic accuracy and saves time.

关键词: Breast cancer, mammography, deep learning, artificial intelligence

Better prognostic determination of cT3 rectal cancer through measurement of distance to mesorectal fascia: A multicenter study

doi: 10.21147/j.issn.1000-9604.2021.05.07

ObjectiveTo forward the magnetic resonance imaging (MRI) based distance between the deepest tumor invasion and mesorectal fascia (DMRF), and to explore its prognosis differentiation value in cT3 stage rectal cancer with comparison of cT3 substage.MethodsThis was a retrospective, multicenter cohort study including cT3 rectal cancer patients undergoing neoadjuvant chemoradiotherapy followed by radical surgery from January 2013 to September 2014. DMRF and cT3 substage were evaluated from baseline MRI. The cutoff of DMRF was determined by disease progression. Multivariate cox regression was used to test the prognostic values of baseline variables. ResultsA total of 804 patients were included, of which 226 (28.1%) developed progression. A DMRF cutoff of 7 mm was chosen. DMRF category, the clock position of the deepest position of tumor invasion (CDTI) and extramural venous invasion (EMVI) were independent predictors for disease progression, and hazard ratios (HRs) were 0.26 [95% confidence interval (95% CI), 0.13−0.56], 1.88 (95% CI, 1.33−2.65) and 1.57 (95% CI, 1.13−2.18), respectively. cT3 substage was not a predictor for disease progression. ConclusionsThe measurement of DMRF value on baseline MRI can better distinguish cT3 rectal cancer prognosis rather than cT3 substage, and was recommended in clinical evaluation.

关键词: Rectal cancer, T3 stage, substage, distance to mesorectal fascia, magnetic resonance imaging

Correlation between imaging features on computed tomography and combined positive score of PD-L1 expression in patients with gastric cancer

doi: 10.21147/j.issn.1000-9604.2022.05.10

ObjectiveTo explore the correlation between computed tomography (CT) features and combined positive score (CPS) of programmed cell death ligand 1 (PD-L1) expression in patients with gastric cancer (GC).MethodsThis study reviewed an institutional database of patients who underwent GC operation without neoadjuvant chemotherapy between December 2019 and September 2020. The CPS results of PD-L1 expression of postoperative histological examination were recorded by pathology. Baseline CT features were measured, and their correlation with CPS 5 or 10 score groups of PD-L1 expression was analyzed.ResultsData for 153 patients with GC were collected. Among them, 124 were advanced GC patients, and 29 were early GC patients. None of the CT features significantly differed between CPS groups with a cutoff score of 5 and a score of 10 in patients with early GC. In advanced GC, the presence of lymph nodes with short diameters >10 mm was significantly different (P=0.024) between the CPS<5 and CPS≥5 groups. CT features such as tumor attenuation in the arterial phase, long and short diameter of the largest lymph node, the sum of long diameter of the two largest lymph nodes, the sum of short diameter of the two largest lymph nodes, and the presence of lymph nodes with short diameters >10 mm significantly differed between the CPS<10 and CPS≥10 groups in advanced GC. The sensitivity, specificity and area under receiver operating characteristic (ROC) curve of logistic regression model for predicting CPS≥10 was 71.7%, 50.0% and 0.671, respectively. Microsatellite instability (MSI) status was significantly different in CPS groups with cutoff score of 5 and 10 in advanced GC patients.ConclusionsCT findings of advanced GC patients with CPS≥10 showed greater arterial phase enhancement and larger lymph nodes. CT has the potential to help screen patients suitable for immunotherapy.

关键词: Gastric cancer, computed tomography, combined positive score

Diffusion-tensor imaging as an adjunct to dynamic contrast-enhanced MRI for improved accuracy of differential diagnosis between breast ductal carcinoma in situ and invasive breast carcinoma

doi: 10.3978/j.issn.1000-9604.2015.03.04

ObjectiveTo determine the value of diffusion-tensor imaging (DTI) as an adjunct to dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) for improved accuracy of differential diagnosis between breast ductal carcinoma in situ (DCIS) and invasive breast carcinoma (IBC).MethodsThe MRI data of 63 patients pathologically confirmed as breast cancer were analyzed. The conventional MRI analysis metrics included enhancement style, initial enhancement characteristic, maximum slope of increase, time to peak, time signal intensity curve (TIC) pattern, and signal intensity on FS-T2WI. The values of apparent diffusion coefficient (ADC), directionally-averaged mean diffusivity (D_avg), exponential attenuation (EA), fractional anisotropy (FA), volume ratio (VR) and relative anisotropy (RA) were calculated and compared between DCIS and IBC. Multivariate logistic regression was used to identify independent factors for distinguishing IBC and DCIS. The diagnostic performance of the diagnosis equation was evaluated using the receiver operating characteristic (ROC) curve. The diagnostic efficacies of DCE-MRI, DWI and DTI were compared independently or combined.ResultsEA value, lesion enhancement style and TIC pattern were identified as independent factor for differential diagnosis of IBC and DCIS. The combination diagnosis showed higher diagnostic efficacy than a single use of DCE-MRI (P=0.02), and the area of the curve was improved from 0.84 (95% CI, 0.67-0.99) to 0.94 (95% CI, 0.85-1.00).ConclusionsQuantitative DTI measurement as an adjunct to DCE-MRI could improve the diagnostic performance of differential diagnosis between DCIS and IBC compared to a single use of DCE-MRI.

关键词: Breast carcinoma in situ, diffusion tensor imaging, magnetic resonance imaging, breast

Prognostic factors for transarterial chemoembolization combined with sustained oxaliplatin-based hepatic arterial infusion chemotherapy of colorectal cancer liver metastasis

doi: 10.21147/j.issn.1000-9604.2017.01.05

ObjectiveTo investigate the prognostic factors in chemorefractory colorectal cancer liver metastasis (CRCLM) patients treated by transarterial chemoembolization (TACE) and sustained hepatic arterial infusion chemotherapy (HAIC). MethodsBetween 2006 and 2015, 162 patients who underwent 763 TACE and HAIC in total were enrolled in this retrospective study, including 110 males and 52 females, with a median age of 60 (range, 26−83) years. Prognostic factors were assessed with Log-rank test, Cox univariate and multivariate analyses. ResultsThe median survival time (MST) and median progression-free survival (PFS) of the 162 patients from first TACE/HAIC were 15.6 months and 5.5 months respectively. Normal serum carbohydrate antigen 19-9 (CA19-9, <37 U/mL) (P<0.001) and carbohydrate antigen 72-4 (CA72-4, <6.7 U/mL) (P=0.026), combination with other local treatment (liver radiotherapy or liver radiofrequency ablation) (P=0.034) and response to TACE/HAIC (P<0.001) were significant factors related to survival after TACE/HAIC in univariate analysis. A multivariate analysis revealed that normal serum CA19-9 (P<0.001), response to TACE/HAIC (P<0.001) and combination with other local treatment (P=0.001) were independent factors among them. ConclusionsOur findings indicate that serum CA19-9 <37 U/mL and response to TACE/HAIC are significant prognostic indicators for this combined treatment, and treated with other local treatment could reach a considerable survival benefit for CRCLM. This could be useful for making decisions regarding the treatment of CRCLM.

关键词: Colorectal cancer, transarterial chemoembolization, hepatic artery infusion chemotherapy

Utility of CT in differentiating liver metastases of well-differentiated gastroenteropancreatic neuroendocrine neoplasms from poorly-differentiated neuroendocrine neoplasms

doi: 10.21147/j.issn.1000-9604.2018.01.04

Objective To determine the capability of dynamic enhanced computed tomography (CT) to differentiate liver metastases (LMs) of well-differentiated from poorly-differentiated gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs). Methods Patients with LMs of GEP-NENs who underwent dynamic enhanced CT examination in Peking University Cancer Hospital from January 2009 to October 2015 were included and data were retrospectively analyzed. We assessed the qualitative and quantitative CT features to identify the significant differentiating CT features of LMs of poorly-differentiated GEP-NENs from those of well-differentiated GEP-NENs using univariate analysis and a multivariate logistic regression model. Results The study included 22 patients with LMs of well-differentiated GEP-NENs and 32 patients with LMs of poorly-differentiated GEP-NENs. Univariate analysis revealed statistically significant differences between the LMs of well- and poorly-differentiated GEP-NENs in terms of feeding arteries (36.4% vs. 75.0%, χ²=8.061, P=0.005), intratumoral neovascularity (18.2% vs. 59.4%, χ²=9.047, P=0.003), lymphadenopathy (27.3% vs. 81.2%, χ²=15.733, P<0.001), tumor-to-aortic ratio in the hepatic arterial and portal venous phase (T-A/AP: 0.297±0.080 vs. 0.251±0.059, t=2.437, P=0.018; T-A/PVP: 0.639±0.138 vs. 0.529±0.117, t=3.163, P=0.003) and tumor-to-liver ratio in the hepatic arterial phase (T-L/AP: 1.108±0.267 vs. 0.907±0.240, t=2.882, P=0.006). The LMs of poorly-differentiated GEP-NENs showed more feeding arteries, more intratumoral neovascularity, more lymphadenopathy and a lower tumor-to-aortic ratio. Multivariate analysis suggested that intratumoral neovascularity [P=0.015, OR=0.108, 95% confidence interval (95% CI), 0.018–0.646], lymphadenopathy (P=0.001, OR=0.055, 95% CI, 0.009–0.323) and T-A/PVP (P=0.004, OR=5.3E–5, 95% CI, 0.000–0.044) were independent factors for differentiating LMs of poorly-differentiated from well-differentiated GEP-NENs. Conclusions Dynamic enhanced CT features (intratumoral neovascularity, lymphadenopathy and T-A/PVP) are useful in the pathological classification of LMs of GEP-NENs.

关键词: Diagnosis, gastroenteropancreatic neuroendocrine neoplasm, neoplasm grading, tomography, X-ray computed

Prediction of pathological nodal stage of locally advanced rectal cancer by collective features of multiple lymph nodes in magnetic resonance images before and after neoadjuvant chemoradiotherapy

doi: 10.21147/j.issn.1000-9604.2019.06.14

ObjectiveTo predict pathological nodal stage of locally advanced rectal cancer by a radiomic method that uses collective features of multiple lymph nodes (LNs) in magnetic resonance images before and after neoadjuvant chemoradiotherapy (NCRT).MethodsA total of 215 patients were included in this study and chronologically divided into the discovery cohort (n=143) and validation cohort (n=72). In total, 2,931 pre-NCRT LNs and 1,520 post-NCRT LNs were delineated from all visible rectal LNs in magnetic resonance images. Geometric, first-order and texture features were extracted from each LN before and after NCRT. Collective features are defined as the maximum, minimum, mean, median value and standard deviation of each feature from all delineated LNs of each participant. LN-model is constructed from collective LN features by logistic regression model with L1 regularization to predict pathological nodal stage (ypN0 or ypN+). Tumor-model is constructed from tumor features for comparison by using DeLong test.ResultsThe LN-model selects 7 features from 412 LN features, and the tumor-model selects 7 features from 82 tumor features. The area under the receiver operating characteristic curve (AUC) of LN-model in the discovery cohort is 0.818 [95% confidence interval (95% CI): 0.745−0.878], significantly (Z=2.09, P=0.037) larger than 0.685 (95% CI: 0.602−0.760) of the tumor-model. The AUC of LN-model in validation cohort is 0.812 (95% CI: 0.703−0.895), significantly (Z=3.106, P=0.002) larger than 0.517 (95% CI: 0.396−0.636) of the tumor-model.ConclusionsThe usage of collective features from all visible rectal LNs performs better than the usage of tumor features for the prediction of pathological nodal stage of locally advanced rectal cancer.

关键词: Lymph node, neoadjuvant therapy, radiomics, rectal cancer

Spectral CT imaging parameters and Ki-67 labeling index in lung adenocarcinoma

doi: 10.21147/j.issn.1000-9604.2020.01.11

ObjectiveTo explore the correlation between the spectral computed tomography (CT) imaging parameters and the Ki-67 labeling index in lung adenocarcinoma.MethodsSpectral CT imaging parameters [iodine concentrations of lesions (ICLs) in the arterial phase (ICLa) and venous phase (ICLv), normalized IC in the aorta (NICa/NICv), slope of the spectral HU curve (λHUa/λHUv) and monochromatic CT number enhancement on 40 keV and 70 keV images (CT40keVa/v, CT70keVa/v)] in 34 lung adenocarcinomas were analyzed, and common molecular markers, including the Ki-67 labeling index, were detected with immunohistochemistry. Different Ki-67 labeling indexes were measured and grouped into four grades according to the number of positive-stained cells (grade 0, ≤1%; 1%<grade 1≤10%; 10%<grade 2≤30%; and grade 3, >30%). One-way analysis of variance (ANOVA) was used to compare the four different grades, and the Bonferroni method was used to correct the P value for multiple comparisons. A Spearman correlation analysis was performed to further research a quantitative correlation between the Ki-67 labeling index and spectral CT imaging parameters.ResultsCT40keVa, CT40keVv, CT70keVa and CT70keVv increased as the grade increased, and CT70keVa and CT70keVv were statistically significant (P<0.05). These four parameters and the Ki-67 labeling index showed a moderate positive correlation with lung adenocarcinoma nodules. ICL, NIC and λHU in the arterial and venous phases were not significantly different among the four grades.ConclusionsThe spectral CT imaging parameters CT40keVa, CT40keVv, CT70keVa and CT70keVv gradually increased with Ki-67 expression and showed a moderate positive correlation with lung adenocarcinomas. Therefore, spectral CT imaging parameter-enhanced monochromatic CT numbers at 70 keV may indicate the extent of proliferation of lung adenocarcinomas.

关键词: Computed tomography, spectral CT, lung adenocarcinoma, Ki-67 labeling index

主管单位：	中国科学院
主办单位：	中国电子学会

版权所有：
通讯地址：
工作邮箱：
办公电话：	020-37627489

Chin J Cancer Res

高级检索