CN110223776B - Lung cancer risk prediction system - Google Patents
Lung cancer risk prediction system Download PDFInfo
- Publication number
- CN110223776B CN110223776B CN201910639611.3A CN201910639611A CN110223776B CN 110223776 B CN110223776 B CN 110223776B CN 201910639611 A CN201910639611 A CN 201910639611A CN 110223776 B CN110223776 B CN 110223776B
- Authority
- CN
- China
- Prior art keywords
- lung cancer
- pulmonary
- nodule
- model
- risk prediction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Investigating Or Analysing Biological Materials (AREA)
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
Abstract
The invention aims to provide a brand-new lung cancer risk prediction system, which comprises an input module, a calculation module and an output module, wherein the input module is used for transmitting patient information to the calculation module, wherein the patient information comprises the age, some characteristics of pulmonary nodules and partial thromboplastin time; the calculation module is internally provided with a lung cancer risk prediction Model1, which is as follows: probability value of lung cancer Y ═ ex/(1+ex) X ═ 3.764+0.063 × age +0.043 × diameter of pulmonary nodules +0.810 × shape +1.641 × texture +0.567 × burr sign-0.042 × partial thromboplastin time; the output module is used for outputting the probability value Y. The method can achieve rapid and effective prediction of lung cancer risk, and has very good application prospect.
Description
Technical Field
The present invention relates to the field of lung cancer diagnostic systems.
Background
Global cancer statistical data shows that in 2018, there are 1810 new cancer cases and 960 ten thousand death cancer cases, wherein the lung cancer accounts for the most of the new cancer cases and the death cases, 11.6 percent and 18.4 percent respectively, and about 1 death occurs in every 5 lung cancer patients.
The survival rate of lung cancer is closely related to the clinical stage of lung cancer at the time of diagnosis. Because early symptoms of lung cancer are not obvious, the lung cancer is diagnosed at an advanced stage, the chance of surgical treatment is lost, and the prognosis is poor.
In order to realize early diagnosis and treatment of lung cancer, low-dose spiral computed tomography (LDCT) lung cancer screening tests have been developed in succession. Although the screening schemes and screening populations of the screening tests are slightly different, most results show that LDCT screening is beneficial to early diagnosis and early intervention of lung cancer and improves the five-year survival rate of lung cancer. However, a higher false positive rate during the screening process is a more problematic issue. In order to ensure that lung cancer patients are diagnosed and treated in time and avoid excessive intervention and unnecessary invasive examination on benign nodule patients, it is important to establish a scientific and normative lung cancer risk assessment model and a corresponding follow-up strategy in the lung nodule diagnosis and treatment process.
Previous studies have shown that the use of high quality risk prediction models can reduce the false positive rate during LDCT lung cancer screening, make screening more efficient, cost effective, and reduce unnecessary detection and invasive procedures. The current lung nodule guideline recommends and well-known models mainly comprise a Mayo model, a Brock model and the like. Both the Brock model and the Mayo model were validated in the Chinese population, but the models were unstable in test performance and had substantially lower AUC than the original study.
Disclosure of Invention
The invention aims to provide new models 1 and 2 for predicting the risk of lung cancer and a lung cancer risk prediction system integrating the models 1 and 2.
The present invention relates to the noun explanation: APTT (activated partial thromboplastin time): partial thrombin activation time;
cea (carcinoembryonic antigen): carcinoembryonic antigen (in blood);
RV/TLC (the ratio of residual volume to total volume capacity): residual capacity lung total ratio.
The "diameter" of the present invention refers to the diameter of a pulmonary nodule, unless otherwise specified.
The technical scheme of the invention comprises the following steps:
a lung cancer risk prediction system comprises an input module, a calculation module and an output module, and is characterized in that:
the input module is used for transmitting the following information of the patient to the calculation module: age, diameter of pulmonary nodules, whether the pulmonary nodule shape is regular, whether the pulmonary nodule texture is pure, whether the pulmonary nodule has burr signs, and partial thromboplastin time;
wherein the diameter of the pulmonary nodule is in mm, and the partial thromboplastin time is in s;
the calculation module is internally provided with a lung cancer risk prediction Model1, which is as follows:
probability value of lung cancer Y ═ ex/(1+ex)；
X ═ 3.764+0.063 × age +0.043 × diameter of pulmonary nodules +0.810 × shape +1.641 × texture +0.567 × burr sign-0.042 × partial thromboplastin time;
if the pulmonary nodule shape is regular, then the "shape" in the formula is 0; otherwise, 1 is taken; the regular refers to the shape of the nodule as a circle or an oval;
if the pulmonary nodule texture is true, then "texture" in the formula is taken to be 0; otherwise, 1 is taken;
if the pulmonary nodule has no burr sign, taking 0 as the burr sign in the formula; otherwise, 1 is taken;
the output module is used for outputting the probability value Y.
As in the foregoing system for predicting the risk of lung cancer, the lung cancer is adenocarcinoma, squamous carcinoma, small cell carcinoma, adenosquamous carcinoma or carcinoid.
The built-in Model1 of the lung cancer risk prediction system is higher than the c-statistics (area under ROC curve) of the existing Mayo Model, Brock Model and PUMC (Peaking Union Medical college) Model, and the built-in Model1 of the invention has higher specificity than another similar Model2, so the system of the invention has excellent prediction capability.
Obviously, many modifications, substitutions, and variations are possible in light of the above teachings of the invention, without departing from the basic technical spirit of the invention, as defined by the following claims.
The foregoing aspects of the present invention are explained in further detail below with reference to specific embodiments. This should not be understood as limiting the scope of the above-described subject matter of the present invention to the following examples. All the technologies realized based on the above contents of the present invention belong to the scope of the present invention.
Drawings
FIG. 1: the resulting Model1 and Model2 ROC curves were plotted based on the set 1 data.
FIG. 2: the resulting Model1 and Model2 ROC curves were plotted based on the set 2 data.
FIG. 3: model1 calibration curve based on set 1 data.
FIG. 4: modle2 calibration curve based on group 1 data.
FIG. 5: clinical decision curve.
FIG. 6: ROC curves for 5 models plotted based on group 1 data.
FIG. 7: ROC curves for 5 models plotted based on group 2 data.
Detailed Description
Early information preparation
1. Basic information: age, history of malignancy (1 for some and 0 for none).
2. Imaging information:
conventional chest CT examinations are performed by one of three multi-detector systems: (a) siemens Medical Systems, Forchheim, Germany; (b) somatom Definition, Siemens Healthcare, Forchheim, Germany; (c) philips Medical Systems, Eindhoven, the Netherlands. The scan parameters were as follows: (a)120kv, 100mAs, rotation speed 0.5s, collimation 16mm × 0.75mm, screw pitch 0.85; (b)120kY, 200mAs, rotation speed 0.33s, collimation 24mm multiplied by 1.2mm, and screw pitch 0.9; (c)120kV, 145mAs, rotation speed 0.42s, collimation 64mm x 0.625mm, pitch 0.891.
The acquired main information includes: lung nodule diameter (unit mm), shape (irregular 1, regular 0), texture (pure ground glass or mixed ground glass 1, pure reality 0), and burr (presence 1, absence 0).
3. Lung function test information: RV/TLC.
4. Blood examination information: APTT (in s), CEA (in ng/ml).
Example 1 method of use of the inventive risk prediction system
After the user inputs the information such as the age, the diameter, the shape, the texture, the burr sign and the APTT through the input module, the calculation module can substitute the information into the Model1 for calculation to obtain the lung cancer ill probability, and the lung cancer ill probability is displayed to the user through the output module.
The model is a key technical characteristic of a lung cancer risk prediction model and is directly related to the prediction effect; the predicted effect of the model of the present invention will be further described below by way of experimental examples.
In the experimental example, another similar prediction Model2 constructed by the inventor is also introduced, which is specifically as follows:
probability value of lung cancer Y ═ ex/(1+ex)；
X-4.367 +0.036 age +0.784 malignancy history +0.072 diameter +0.711 shape +1.770 texture +1.015 Burr sign +0.105 carcinoembryonic antigen concentration in blood +0.011 residual lung volume ratio;
if no malignant tumor history exists, taking 0 as the malignant tumor history in the formula; otherwise, 1 is taken;
if the pulmonary nodule shape is regular, then the "shape" in the formula is 0; otherwise, 1 is taken; the regular refers to the shape of the nodule as a circle or an oval;
if the pulmonary nodule texture is true, then "texture" in the formula is taken to be 0; otherwise, 1 is taken;
if the pulmonary nodule has no burr sign, taking 0 as the burr sign in the formula; otherwise, 1 is taken;
the output module is used for outputting the probability value Y.
Experimental example 1 verification of the Effect of the model of the present invention
1. Inclusion criteria
This section is included in 2010-2017 patients with benign and malignant pulmonary nodules diagnosed in western China Hospital, Sichuan university. The major inclusion criteria are as follows: (1) chest CT shows focal, roundlike, dense or hypo-solid pulmonary nodules with a diameter of 5-30 mm; (2) the pathological diagnosis of the nodules is clear. The major exclusion criteria were as follows: (1) the mediastinal window shows calcified nodules; (2) definitive diagnosis of multiple primary lung cancer or multiple benign nodules; (3) pulmonary metastases; (4) with atelectasis, enlargement of the pulmonary lymph nodes or pleural effusion. The study was approved by the ethical committee of the institution.
2. Study object
2821 lung nodule patients with definite pathological diagnosis are included in the study, wherein 1813 cases are primary lung cancer, mainly including adenocarcinoma (1685), squamous carcinoma (105), small cell carcinoma (9), adenosquamous carcinoma (6), carcinoid (4) and the like (Table 1A); 1008 cases are benign lesions, mainly including pneumonia nodules (528), benign lung tumors (237), tuberculosis (178), fungal infection (8), and lymph node hyperplasia (7) (table 1B). All patients had a male to female ratio of 0.8: 1.0(1270/1551), mean age 56 years, and mean nodule diameter 18 mm.
TABLE 1 pathological type distribution of study subjects
A. Distribution of malignant nodule pathology types
B. Distribution of benign nodal pathology types
Subjects were randomized into 2 groups: group 1 had 1880 cases, including 669 benign nodules and 1211 malignant nodules; the remainder is set 2.
In group 1, lung cancer patients were older (59vs 51 years), female more (58.0% vs 50.8%), more with a history of malignancy and a family history of malignancy (P < 0.05) than benign nodules; chest CT showed larger diameter of malignant nodules (18vs 17mm), multiple superior lung lobes (60.3% vs 49.0%), more irregular shape (80.4% vs 58.0%), multiple burr features (56.6% vs 36.0%) and lobular features (54.85vs 41.1%), and more ground glass nodules (18.2% vs 5.8% for pure ground glass nodules; 19.8% vs 13.0% for mixed ground glass nodules); laboratory tests show that lung cancer patients have low red blood cell numbers (4.54vs 4.62X 10^12/L), shortened partial thromboplastin time (26.94vs 28.00s), higher tumor markers CEA (4.09vs 2.01ng/ml) and CYFRA21-1(2.23vs 2.04 ng/ml); pulmonary function examination showed that lung cancer patients had lower FEV1/FVC (78.02% vs 79.37%), MMEF% (71.54% vs 76.15%), and V50% (77.14% vs 83.42%), and higher RV/TLC (41.45% vs 37.25%).
In group 2, the clinical data distribution of benign nodule patients and lung cancer patients was substantially similar to that of group 1, but different from each other. The smoking rate in lung cancer patients was lower (26.9% vs 40.1%), there was no history of overt malignancy or family history of malignancy (P > 0.05), the prothrombin time was slightly longer (11.22vs 11.01 s); furthermore, there was no significant difference in FEV1/FVC between lung cancer patients and benign nodule patients (78.81% vs 79.91%), but lung cancer patients had lower V25% (60.75% vs 69.10%). 3. Model validation
3.1 degree of model identification
ROC curves for Model1 and Model2 were plotted for cohort 1 and cohort 2, respectively, and Model c-statistics was calculated. It can be seen from FIGS. 1 and 2 that Model1 and Model2 have almost the same AUC, with c-statistics of 0.78(0.76-0.80) and 0.78(0.75-0.82), respectively, in cohort 1; in group 2, the c-statistics were 0.76(0.72-0.79) and 0.72(0.66-0.78), respectively.
Next, the sensitivities (71% vs 80%; specificity (74% vs 64%), positive predictive values (83% vs 85%), negative predictive values (58% vs 53%), positive likelihood ratios (2.67vs2.05), negative likelihood ratios (2.54vs 3.10), and accuracies (72% vs 75%) of Model1 and Model2, respectively, were calculated in cohort 1. Model1 was found to be slightly less sensitive but more specific than Model2 for Model 1.
Finally, the sensitivity (83% vs 80%; specificity (59% vs 56%), positive predictive value (77% vs 86%), negative predictive value (66% vs 46%), positive likelihood ratio (1.94vs1.78), negative likelihood ratio (3.41vs 2.93), and accuracy (74% vs 75%) were also calculated for Model1 and Model2 in cohort 2. The specificity of Model1 and Model2 was slightly reduced. In general, the accuracy of Model2 was higher than that of Model1 (Table 2) in either cohort 1 or cohort 2.
TABLE 2Model1 and Model2 degrees of identification
Note: AUC, area under the curve, AUC
3.2 model calibration
In order to detect whether the prediction result of the Model is consistent with the actual result, respectively drawing Model1 and Model2 calibration curves in group 1, as shown in fig. 3 and 4, respectively, wherein application in the graphs shows a fitting line of the predicted value corresponding to the actual value when the deviation is not corrected, and Bias-corrected display shows the fitting line after the deviation is corrected; ideal represents the most Ideal relationship between the predicted probability and the actual probability.
It can be seen that the predicted probability of the model substantially matches the actual probability.
3.3 model clinical benefit
To assess the clinical utility value of the model, a clinical decision curve was plotted in group 1 (fig. 5). The graph shows the clinical net benefit of the model (ordinate) at different cut-off points (abscissa). Model and Model2 have a cutoff of 0.66 based on the john index, with some clinical benefit.
4. Comparison with published models
The study screened Mayo Model, Brock Model and PUMC Model, evaluated the test performance in groups 1 and 2, respectively, and compared to the newly established models Model1 and Model 2. Three model parameters have been published as follows:
mayo Model: x ═ 6.8272+ (0.0391 × age) + (0.7917 × smoking) + (1.3388 × history of malignancy) + (0.1274 × diameter) + (1.0407 × burred) + (0.7838 × upper leaf); note: nodule diameter in mm
Brock Model (select Model1b, simplified Model includes spike characterization): x ═ 6.6144+0.6467 × sex-5.553 × diameter +0.6009 × site +0.9309 × burr; note: gender "female" was 1, mm in diameter and was converted as follows:
peking Union Medical College model (PUMC model): x ═ 4.294+ (0.035 × age) + (0.221 × CEA) + (0.200 × CYFRA 21-1) + (1.029 × smoking) + (0.974 × family history of malignancy) + (0.633 × diameter) + (-1.631 × clear border) + (-1.923 × satellite foci) + (2.673 × leaf syndrome) + (-3.295 × calcified) + (2.027 × burr syndrome); note: the nodule diameter unit cm does not include the index of 'satellite focus' and 'calcification' when calculating the probability.
Comparison of c-statistics of each large Model revealed that Moldel1 and Model2 were superior to the other three large models, whether in group 1 or group 2. In group 1, the c-statistics of the Mayo model, the Brock model and the PUMC model were 0.63(0.61-0.66), 0.59(0.57-0.62) and 0.65(0.61-0.68), respectively; in group 2, the c-statistics were 0.62(0.59-0.66), 0.62(0.58-0.66) and 0.63(0.58-0.69) in this order (Table 3, FIG. 6, FIG. 7).
Table 3 comparison of the models
Note:*validation Brock Model1 b;#does not include the indexes of ' satellite range ' and ' calcification
5. Conclusion
The Model1 built in the lung cancer risk prediction system has stronger specificity than that of a similar Model 2; the Model1 built in the lung cancer risk prediction system of the invention is also higher than the c-statistics of the existing Mayo Model, Brock Model and PUMC Model.
The risk prediction system of the invention is provided with the lung cancer risk prediction Model1, so that the risk of lung cancer can be predicted very effectively, and the application prospect is good.
Claims (2)
1. A lung cancer risk prediction system comprises an input module, a calculation module and an output module, and is characterized in that:
the input module is used for transmitting the following information of the patient to the calculation module: age, diameter of pulmonary nodules, whether the pulmonary nodule shape is regular, whether the pulmonary nodule texture is pure, whether the pulmonary nodule has burr signs, and partial thromboplastin time;
wherein the diameter of the pulmonary nodule is in mm, and the partial thromboplastin time is in s;
the calculation module is internally provided with a lung cancer risk prediction Model1, which is as follows:
probability value of lung cancer Y ═ ex/(1+ex)；
X ═ 3.764+0.063 × age +0.043 × diameter of pulmonary nodules +0.810 × shape +1.641 × texture +0.567 × burr sign-0.042 × partial thromboplastin time;
if the pulmonary nodule shape is regular, then the "shape" in the formula is 0; otherwise, 1 is taken; the regular refers to the shape of the nodule as a circle or an oval;
if the pulmonary nodule texture is true, then "texture" in the formula is taken to be 0; otherwise, 1 is taken;
if the pulmonary nodule has no burr sign, taking 0 as the burr sign in the formula; otherwise, 1 is taken;
the output module is used for outputting the probability value Y.
2. The lung cancer risk prediction system of claim 1, wherein the lung cancer is adenocarcinoma, squamous carcinoma, small cell carcinoma, adenosquamous carcinoma, or carcinoid.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910639611.3A CN110223776B (en) | 2019-07-15 | 2019-07-15 | Lung cancer risk prediction system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910639611.3A CN110223776B (en) | 2019-07-15 | 2019-07-15 | Lung cancer risk prediction system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110223776A CN110223776A (en) | 2019-09-10 |
| CN110223776B true CN110223776B (en) | 2021-01-01 |
Family
ID=67813442
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910639611.3A Active CN110223776B (en) | 2019-07-15 | 2019-07-15 | Lung cancer risk prediction system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110223776B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105717147A (en) * | 2016-03-29 | 2016-06-29 | 复旦大学附属中山医院 | Model for predicating lung cancer risks for people suffering from pulmonary nodules among China urban population on basis of CT (computed tomography) images and biomarker spectrums |
| CN105717146A (en) * | 2016-03-29 | 2016-06-29 | 复旦大学附属中山医院 | Kit for predicating lung cancer risk for high-risk groups among China urban population on basis of CT (computed tomography) images and biomarker spectrums |
| CN105891482A (en) * | 2016-03-29 | 2016-08-24 | 复旦大学附属中山医院 | Biomarker spectrum-based lung cancer risk prediction model for Chinese rural residues suffering from pulmonary nodules |
| CN107292114A (en) * | 2017-06-28 | 2017-10-24 | 中日友好医院 | A kind of method for building up of the pernicious Probabilistic Prediction Model of solitary pulmonary nodule |
| CN108520778A (en) * | 2018-04-20 | 2018-09-11 | 中日友好医院 | A kind of screening lung cancer system and method based on big data |
| CN109817336A (en) * | 2019-01-21 | 2019-05-28 | 杭州英库医疗科技有限公司 | A kind of early stage of lung cancer risk evaluation model method for building up of combination imaging Analysis result |
-
2019
- 2019-07-15 CN CN201910639611.3A patent/CN110223776B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105717147A (en) * | 2016-03-29 | 2016-06-29 | 复旦大学附属中山医院 | Model for predicating lung cancer risks for people suffering from pulmonary nodules among China urban population on basis of CT (computed tomography) images and biomarker spectrums |
| CN105717146A (en) * | 2016-03-29 | 2016-06-29 | 复旦大学附属中山医院 | Kit for predicating lung cancer risk for high-risk groups among China urban population on basis of CT (computed tomography) images and biomarker spectrums |
| CN105891482A (en) * | 2016-03-29 | 2016-08-24 | 复旦大学附属中山医院 | Biomarker spectrum-based lung cancer risk prediction model for Chinese rural residues suffering from pulmonary nodules |
| CN107292114A (en) * | 2017-06-28 | 2017-10-24 | 中日友好医院 | A kind of method for building up of the pernicious Probabilistic Prediction Model of solitary pulmonary nodule |
| CN108520778A (en) * | 2018-04-20 | 2018-09-11 | 中日友好医院 | A kind of screening lung cancer system and method based on big data |
| CN109817336A (en) * | 2019-01-21 | 2019-05-28 | 杭州英库医疗科技有限公司 | A kind of early stage of lung cancer risk evaluation model method for building up of combination imaging Analysis result |
Non-Patent Citations (2)
| Title |
|---|
| 张满.孤立性肺结节良恶性预测模型的建立.《中国博士学位论文全文数据库 医药卫生科技辑 》.2017,E072-94. * |
| 曹圆圆 等.缩短的APTT水平与肺癌患者病情及预后的关系.《医学研究杂志》.2016,第45卷(第11期),第129-132页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110223776A (en) | 2019-09-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Maida et al. | Screening of colorectal cancer: present and future | |
| Chew et al. | Critical analysis of mucin and signet ring cell as prognostic factors in an Asian population of 2,764 sporadic colorectal cancers | |
| CN110223775B (en) | Lung cancer risk prediction system | |
| Sun et al. | Radiomics study for predicting the expression of PD-L1 in non-small cell lung cancer based on CT images and clinicopathologic features | |
| Lee et al. | Usefulness of texture analysis in differentiating transient from persistent part-solid nodules (PSNs): a retrospective study | |
| Zhang et al. | Differentiation combined hepatocellular and cholangiocarcinoma from intrahepatic cholangiocarcinoma based on radiomics machine learning | |
| Zhang et al. | Development and validation of MRI‐based deep learning models for prediction of microsatellite instability in rectal cancer | |
| Alfonso et al. | Screening for occult malignancy with FDG‐PET/CT in patients with unprovoked venous thromboembolism | |
| Yang et al. | Predictive model for neoplastic potential of gallbladder polyp | |
| Liu et al. | A noninvasive multianalytical approach for lung cancer diagnosis of patients with pulmonary nodules | |
| Sun et al. | Preoperative prediction for Lauren type of gastric cancer: A radiomics nomogram analysis based on CT images and clinical features | |
| Huang et al. | CT-based radiomics combined with signs: a valuable tool to help radiologist discriminate COVID-19 and influenza pneumonia | |
| Gurbani et al. | Evaluation of radiomics and machine learning in identification of aggressive tumor features in renal cell carcinoma (RCC) | |
| Zheng et al. | A CT-based radiomics signature for preoperative discrimination between high and low expression of programmed death ligand 1 in head and neck squamous cell carcinoma | |
| He et al. | Molecular diagnosis of pancreatobiliary tract cancer by detecting mutations and methylation changes in bile samples | |
| Inchingolo et al. | Radiomics in colorectal cancer patients | |
| Huang et al. | The prognostic significance of pure ground glass opacities in lung cancer computed tomographic images | |
| Yang et al. | A case report of primary signet ring cell carcinoma of the lung: imaging study and literature review | |
| CN110223776B (en) | Lung cancer risk prediction system | |
| Sun et al. | Radiomics nomogram for the prediction of Ki-67 index in advanced non-small cell lung cancer based on dual-phase enhanced computed tomography | |
| Budai et al. | Radiomics analysis of contrast-enhanced CT scans can distinguish between clear cell and non-clear cell renal cell carcinoma in different imaging protocols | |
| Tian et al. | Differential diagnostic value of histology in MPLC and IPM: a systematic review and meta-analysis | |
| Dai et al. | Predictive features of thymic carcinoma and high-risk thymomas using random forest analysis | |
| Pelaez-Luna et al. | Cyst fluid analysis to diagnose pancreatic cystic lesions: an as yet unfulfilled promise | |
| Gao et al. | Radiomics study on pulmonary infarction mimicking community‐acquired pneumonia |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |