HRCT morphological characteristics distinguishing minimally invasive pulmonary adenocarcinoma from invasive pulmonary adenocarcinoma appearing as subsolid nodules with a diameter of ≤3 cm

Evaluation of the invasiveness of pure ground-glass nodules based on dual-head ResNet technique

OBJECTIVE

To intelligently evaluate the invasiveness of pure ground-glass nodules with multiple classifications using deep learning.

METHODS

pGGNs in 1136 patients were pathologically confirmed as lung precursor lesions [atypical adenomatous hyperplasia (AAH) and adenocarcinoma in situ (AIS)], minimally invasive adenocarcinoma (MIA), or invasive adenocarcinoma (IAC). Four different models [EfficientNet-b0 2D, dual-head ResNet_3D, a 3D model combining three features (3D_3F), and a 3D model combining 19 features (3D_19F)] were constructed to evaluate the invasiveness of pGGNs using the EfficientNet and ResNet networks. The Obuchowski index was used to evaluate the differences in diagnostic efficiency among the four models.

RESULTS

The patients with pGGNs (360 men, 776 women; mean age, 54.63 ± 12.36 years) included 235 cases of AAH + AIS, 332 cases of MIA, and 569 cases of IAC. In the validation group, the areas under the curve in detecting the invasiveness of pGGNs as a three-category classification (AAH + AIS, MIA, IAC) were 0.8008, 0.8090, 0.8165, and 0.8158 for EfficientNet-b0 2D, dual-head ResNet_3D, 3D_3F, and 3D_19F, respectively, whereas the accuracies were 0.6422, 0.6158, 0.651, and 0.6364, respectively. The Obuchowski index revealed no significant differences in the diagnostic performance of the four models.

CONCLUSIONS

The dual-head ResNet_3D_3F model had the highest diagnostic efficiency for evaluating the invasiveness of pGGNs in the four models.

Estimation of pathological subtypes in subsolid lung nodules using artificial intelligence

Objective

This study aimed to investigate the value of artificial intelligence (AI) for distinguishing pathological subtypes of invasive pulmonary adenocarcinomas in patients with subsolid nodules (SSNs).

Materials and methods

This retrospective study included 110 consecutive patients with 120 SSNs. The qualitative and quantitative imaging characteristics of SSNs were extracted automatically using an artificially intelligent assessment system. Then, radiologists had to verify these characteristics again. We split all cases into two groups: non-IA including 11 Atypical adenomatous hyperplasia (AAH) and 25 adenocarcinoma in situ (AIS) or IA including 7 minimally invasive adenocarcinoma (MIA) and 77 invasive adenocarcinoma (IAC). Variables that exhibited statistically significant differences between the non-IA and IA in the univariate analysis were included in the multivariate logistic regression analysis. Receiver operating characteristic (ROC) analyses were conducted to determine the cut-off values and their diagnostic performances.

Results

Multivariate logistic regression analysis showed that the major diameter (odds ratio [OR] = 1.38; 95 % confidence interval [CI], 1.02-1.87; P = 0.036) and entropy of three-dimensional(3D) CT value (OR = 3.73, 95 % CI, 1.13-2.33, P = 0.031) were independent risk factors for adenocarcinomas. The cut-off values of the major diameter and the entropy of 3D CT value for the diagnosis of invasive adenocarcinoma were 15.5 mm and 5.17, respectively. To improve the classification performance, we fused the major diameter and the entropy of 3D CT value as a combined model, and the (AUC） of the model was 0.868 (sensitivity = 0.845, specificity = 0.806).

Conclusion

The major diameter and entropy of 3D CT value can distinguish non-IA from IA. AI can improve performance in distinguishing pathological subtypes of invasive pulmonary adenocarcinomas in patients with SSNs.

A Radiological-Radiomics model for differentiation between minimally invasive adenocarcinoma and invasive adenocarcinoma less than or equal to 3 cm: A two-center retrospective study

OBJECTIVE

To develop a Radiological-Radiomics (R-R) combined model for differentiation between minimal invasive adenocarcinoma (MIA) and invasive adenocarcinoma (IA) of lung adenocarcinoma (LUAD) and evaluate its predictive performance.

METHODS

The clinical, pathological, and imaging data of a total of 509 patients (522 lesions) with LUAD diagnosed by surgical pathology from 2 medical centres were retrospectively collected, with 392 patients (402 lesions) from center 1 trained and validated using a five-fold cross-validation method, and 117 patients (120 lesions) from center 2 serving as an independent external test set. The least absolute shrinkage and selection operator (LASSO) method was utilized to filter features. Logistic regression was used to construct three models for predicting IA, namely, Radiological model, Radiomics model, and R-R model. Also, receiver operating curve curves (ROCs) were plotted, generating corresponding area under the curve (AUC), sensitivity, specificity, and accuracy.

RESULTS

The R-R model for IA prediction achieved an AUC of 0.918 (95 % CI: 0.889-0.947), a sensitivity of 80.3 %, a specificity of 88.2 %, and an accuracy of 82.1 % in the training set. In the validation set, this model exhibited an AUC of 0.906 (95 % CI: 0.842-0.970), a sensitivity of 79.9 %, a specificity of 88.1 %, and an accuracy of 81.8 %. In the external test set, the AUC was 0.894 (95 % CI: 0.824-0.964), a sensitivity of 84.8 %, a specificity of 78.6 %, and an accuracy of 83.3 %.

CONCLUSION

The R-R model showed excellent diagnostic performance in differentiating MIA and IA, which can provide a certain reference for clinical diagnosis and surgical treatment plans.

A nomogram based on the quantitative and qualitative features of CT imaging for the prediction of the invasiveness of ground glass nodules in lung adenocarcinoma

PURPOSE

Based on the quantitative and qualitative features of CT imaging, a model for predicting the invasiveness of ground-glass nodules (GGNs) was constructed, which could provide a reference value for preoperative planning of GGN patients.

MATERIALS AND METHODS

Altogether, 702 patients with GGNs (including 748 GGNs) were included in this study. The GGNs operated between September 2020 and July 2022 were classified into the training group (n = 555), and those operated between August 2022 and November 2022 were classified into the validation group (n = 193). Clinical data and the quantitative and qualitative features of CT imaging were harvested from these patients. In the training group, the quantitative and qualitative characteristics in CT imaging of GGNs were analyzed by using performing univariate and multivariate logistic regression analyses, followed by constructing a nomogram prediction model. The differentiation, calibration, and clinical practicability in both the training and validation groups were assessed by the nomogram models.

RESULTS

In the training group, multivariate logistic regression analysis disclosed that the maximum diameter (OR = 4.707, 95%CI: 2.06-10.758), consolidation/tumor ratio (CTR) (OR = 1.027, 95%CI: 1.011-1.043), maximum CT value (OR = 1.025, 95%CI: 1.004-1.047), mean CT value (OR = 1.035, 95%CI: 1.008-1.063; P = 0.012), spiculation sign (OR = 2.055, 95%CI: 1.148-3.679), and vascular convergence sign (OR = 2.508, 95%CI: 1.345-4.676) were independent risk parameters for invasive adenocarcinoma. Based on these findings, we established a nomogram model for predicting the invasiveness of GGN, and the AUC was 0.910 (95%CI: 0.885-0.934) and 0.902 (95%CI: 0.859-0.944) in the training group and the validation group, respectively. The internal validation of the Bootstrap method showed an AUC value of 0.905, indicating a good differentiation of the model. Hosmer-Lemeshow goodness of fit test for the training and validation groups indicated that the model had a good fitting effect (P > 0.05). Furthermore, the calibration curve and decision analysis curve of the training and validation groups reflected that the model had a good calibration degree and clinical practicability.

CONCLUSION

Combined with the quantitative and qualitative features of CT imaging, a nomogram prediction model can be created to forecast the invasiveness of GGNs. This model has good prediction efficacy for the invasiveness of GGNs and can provide help for the clinical management and decision-making of GGNs.

The value of CT shape quantification in predicting pathological classification of lung adenocarcinoma

OBJECTIVE

To evaluate whether quantification of lung GGN shape is useful in predicting pathological categorization of lung adenocarcinoma and guiding the clinic.

METHODS

98 patients with primary lung adenocarcinoma were pathologically confirmed and CT was performed preoperatively, and all lesions were pathologically ≤ 30 mm in size. On CT images, we measured the maximum area of the lesion's cross-section (MA). The longest diameter of the tumor (LD) was marked with points A and B, and the perpendicular diameter (PD) was marked with points C and D, which was the longest diameter perpendicular to AB. and D, which was the longest diameter perpendicular to AB. We took angles A and B as big angle A (BiA) and small angle A (SmA). We measured the MA, LD, and PD, and for analysis we derived the LD/PD ratio and the BiA/SmA ratio. The data were analysed using the chi-square test, t-test, ROC analysis, and binary logistic regression analysis.

RESULTS

Precursor glandular lesions (PGL) and microinvasive adenocarcinoma (MIA) were distinguished from invasive adenocarcinoma (IAC) by the BiA/SmA ratio and LD, two independent factors (p = 0.007, p = 0.018). Lung adenocarcinoma pathological categorization was indicated by the BiA/SmA ratio of 1.35 and the LD of 11.56 mm with sensitivity of 81.36% and 71.79%, respectively; specificity of 71.79% and 74.36%, respectively; and AUC of 0.8357 (95% CI: 0.7558-0.9157, p < 0.001), 0.8666 (95% CI: 0.7866-0.9465, p < 0.001), respectively. In predicting the pathological categorization of lung adenocarcinoma, the area under the ROC curve of the BiA/SmA ratio combined with LD was 0.9231 (95% CI: 0.8700-0.9762, p < 0.001), with a sensitivity of 81.36% and a specificity of 89.74%.

CONCLUSIONS

Quantification of lung GGN morphology by the BiA/SmA ratio combined with LD could be helpful in predicting pathological classification of lung adenocarcinoma.

Artificial-intelligence-based computed tomography histogram analysis predicting tumor invasiveness of lung adenocarcinomas manifesting as radiological part-solid nodules

Background

Tumor invasiveness plays a key role in determining surgical strategy and patient prognosis in clinical practice. The study aimed to explore artificial-intelligence-based computed tomography (CT) histogram indicators significantly related to the invasion status of lung adenocarcinoma appearing as part-solid nodules (PSNs), and to construct radiomics models for prediction of tumor invasiveness.

Methods

We identified surgically resected lung adenocarcinomas manifesting as PSNs in Peking University People's Hospital from January 2014 to October 2019. Tumors were categorized as adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA), and invasive adenocarcinoma (IAC) by comprehensive pathological assessment. The whole cohort was randomly assigned into a training (70%, n=832) and a validation cohort (30%, n=356) to establish and validate the prediction model. An artificial-intelligence-based algorithm (InferRead CT Lung) was applied to extract CT histogram parameters for each pulmonary nodule. For feature selection, multivariate regression models were built to identify factors associated with tumor invasiveness. Logistic regression classifier was used for radiomics model building. The predictive performance of the model was then evaluated by ROC and calibration curves.

Results

In total, 299 AIS/MIAs and 889 IACs were included. In the training cohort, multivariate logistic regression analysis demonstrated that age [odds ratio (OR), 1.020; 95% CI, 1.004-1.037; p=0.017], smoking history (OR, 1.846; 95% CI, 1.058-3.221; p=0.031), solid mean density (OR, 1.014; 95% CI, 1.004-1.024; p=0.008], solid volume (OR, 5.858; 95% CI, 1.259-27.247; p = 0.037), pleural retraction sign (OR, 3.179; 95% CI, 1.057-9.559; p = 0.039), variance (OR, 0.570; 95% CI, 0.399-0.813; p=0.002), and entropy (OR, 4.606; 95% CI, 2.750-7.717; p<0.001) were independent predictors for IAC. The areas under the curve (AUCs) in the training and validation cohorts indicated a better discriminative ability of the histogram model (AUC=0.892) compared with the clinical model (AUC=0.852) and integrated model (AUC=0.886).

Conclusion

We developed an AI-based histogram model, which could reliably predict tumor invasiveness in lung adenocarcinoma manifesting as PSNs. This finding would provide promising value in guiding the precision management of PSNs in the daily practice.

Value of preoperative 18F-FDG PET/CT and HRCT in predicting the differentiation degree of lung adenocarcinoma dominated by solid density

Purpose

To evaluate the value of positron emission tomography/computed tomography (PET/CT) combined with high-resolution CT (HRCT) in determining the degree of differentiation of lung adenocarcinoma.

Methods

From January 2018 to January 2022, 88 patients with solid density nodules that are lung adenocarcinoma were surgically treated. All patients were examined using HRCT and PET/CT before surgery. During HRCT, two independent observers assessed the presence of lobulation, spiculation, pleural indentation, vascular convergence, and air bronchial signs (bronchial distortion and bronchial disruption). The diameter and CT value of the nodules were measured simultaneously. During PET/CT, the maximum standard uptake value (SUVmax), mean standard uptake value (SUVmean), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) of the nodules were measured. The risk factors of pathological classification were predicted by logistic regression analysis.

Results

All 88 patients (mean age 60 ± 8 years; 44 males and 44 females) were evaluated. The average nodule size was 2.6 ± 1.1 cm. The univariate analysis showed that carcinoembryonic antigen (CEA), pleural indentation, vascular convergence, bronchial distortion, and higher SUVmax were more common in poor differentiated lung adenocarcinoma, and in the multivariate analysis, pleural indentation, vascular convergence, and SUVmax were predictive factors. The combined diagnosis using these three factors showed that the area under the curve (AUC) was 0.735.

Conclusion

SUVmax >6.99 combined with HRCT (pleural indentation sign and vascular convergence sign) is helpful to predict the differentiation degree of lung adenocarcinoma dominated by solid density.

A radiomics nomogram for invasiveness prediction in lung adenocarcinoma manifesting as part-solid nodules with solid components smaller than 6 mm

Objective

To investigate whether radiomics can help radiologists and thoracic surgeons accurately predict invasive adenocarcinoma (IAC) manifesting as part-solid nodules (PSNs) with solid components <6 mm and provide a basis for rational clinical decision-making.

Materials and Methods

In total, 1,210 patients (mean age ± standard deviation: 54.28 ± 11.38 years, 374 men and 836 women) from our hospital and another hospital with 1,248 PSNs pathologically diagnosed with adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA), or IAC were enrolled in this study. Among them, 1,050 cases from our hospital were randomly divided into a derivation set (n = 735) and an internal validation set (n = 315), 198 cases from another hospital were used for external validation. Each labeled nodule was segmented, and 105 radiomics features were extracted. Least absolute shrinkage and selection operator (LASSO) was used to calculate Rad-score and build the radiomics model. Multivariable logistic regression was conducted to identify the clinicoradiological predictors and establish the clinical-radiographic model. The combined model and predictive nomogram were developed based on identified clinicoradiological independent predictors and Rad-score using multivariable logistic regression analysis. The predictive performances of the three models were compared via receiver operating characteristic (ROC) curve analysis. Decision curve analysis (DCA) was performed on both the internal and external validation sets to evaluate the clinical utility of the nomogram.

Results

The radiomics model showed superior predictive performance than the clinical-radiographic model in both internal and external validation sets (Az values, 0.884 vs. 0.810, p = 0.001; 0.924 vs. 0.855, p < 0.001, respectively). The combined model showed comparable predictive performance to the radiomics model (Az values, 0.887 vs. 0.884, p = 0.398; 0.917 vs. 0.924, p = 0.271, respectively). The clinical application value of the nomogram developed based on the Rad-score, maximum diameter, and lesion shape was confirmed, and DCA demonstrated that application of the Rad-score would be beneficial for radiologists predicting invasive lesions.

Conclusions

Radiomics has the potential as an independent diagnostic tool to predict the invasiveness of PSNs with solid components <6 mm.

A CT-based deep learning model for subsolid pulmonary nodules to distinguish minimally invasive adenocarcinoma and invasive adenocarcinoma

OBJECTIVE

To develop and validate a deep learning nomogram (DLN) model constructed from non-contrast computed tomography (CT) images for discriminating minimally invasive adenocarcinoma (MIA) from invasive adenocarcinoma (IAC) in patients with subsolid pulmonary nodules (SSPNs).

MATERIALS AND METHODS

In total, 365 consecutive patients who presented with SSPNs and were pathologically diagnosed with MIA or IAC after surgery, were recruited from two medical institutions from 2016 to 2019. Deep learning features were selected from preoperative CT images using convolutional neural network. Deep learning signature (DLS) was developed via the least absolute shrinkage and selection operator (LASSO). New DLN integrating clinical variables, subjective CT findings, and DLS was constructed. The diagnostic efficiency and discriminative capability were analyzed using the receiver operating characteristic method and decision curve analysis (DCA).

RESULTS

In total, 18 deep learning features with non-zero coefficients were enrolled to develop the DLS, which was statistically different between the MIA and IAC groups. Independent predictors of DLS and lobulated sharp were used to build the DLN. The areas under the curves of the DLN were 0.889 (95% confidence interval (CI): 0.824-0.936), 0.915 (95% CI: 0.846-0.959), and 0.914 (95% CI: 0.848-0.958) in the training, internal validation, and external validation cohorts, respectively. After stratification analysis and DCA, the DLN showed potential generalization ability.

CONCLUSION

The DLN incorporating the DLS and subjective CT findings have strong potential to distinguish MIA from IAC in patients with SSPNs, and will facilitate the suitable treatment method selection for the management of SSPNs.

Determining the invasiveness of ground-glass nodules using a 3D multi-task network

OBJECTIVES

The aim of this study was to determine the invasiveness of ground-glass nodules (GGNs) using a 3D multi-task deep learning network.

METHODS

We propose a novel architecture based on 3D multi-task learning to determine the invasiveness of GGNs. In total, 770 patients with 909 GGNs who underwent lung CT scans were enrolled. The patients were divided into the training (n = 626) and test sets (n = 144). In the test set, invasiveness was classified using deep learning into three categories: atypical adenomatous hyperplasia (AAH) and adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA), and invasive pulmonary adenocarcinoma (IA). Furthermore, binary classifications (AAH/AIS/MIA vs. IA) were made by two thoracic radiologists and compared with the deep learning results.

RESULTS

In the three-category classification task, the sensitivity, specificity, and accuracy were 65.41%, 82.21%, and 64.9%, respectively. In the binary classification task, the sensitivity, specificity, accuracy, and area under the ROC curve (AUC) values were 69.57%, 95.24%, 87.42%, and 0.89, respectively. In the visual assessment of GGN invasiveness of binary classification by the two thoracic radiologists, the sensitivity, specificity, and accuracy of the senior and junior radiologists were 58.93%, 90.51%, and 81.35% and 76.79%, 55.47%, and 61.66%, respectively.

CONCLUSIONS

The proposed multi-task deep learning model achieved good classification results in determining the invasiveness of GGNs. This model may help to select patients with invasive lesions who need surgery and the proper surgical methods.

KEY POINTS

• The proposed multi-task model has achieved good classification results for the invasiveness of GGNs. • The proposed network includes a classification and segmentation branch to learn global and regional features, respectively. • The multi-task model could assist doctors in selecting patients with invasive lesions who need surgery and choosing appropriate surgical methods.

Deep learning-based differentiation of invasive adenocarcinomas from preinvasive or minimally invasive lesions among pulmonary subsolid nodules

OBJECTIVES

To evaluate a deep learning-based model using model-generated segmentation masks to differentiate invasive pulmonary adenocarcinoma (IPA) from preinvasive lesions or minimally invasive adenocarcinoma (MIA) on CT, making comparisons with radiologist-derived measurements of solid portion size.

METHODS

Four hundred eleven subsolid nodules (SSNs) (120 preinvasive lesions or MIAs and 291 IPAs) in 333 patients who underwent surgery between June 2010 and August 2016 were retrospectively included to develop the model (370 SSNs in 293 patients for training and 41 SSNs in 40 patients for tuning). Ninety SSNs of 2 cm or smaller (45 preinvasive lesions or MIAs and 45 IPAs) resected in 2018 formed a validation set. Six radiologists measured the solid portion of each nodule. Performances of the model and radiologists were assessed using receiver operating characteristics curve analysis.

RESULTS

The deep learning model differentiated IPA from preinvasive lesions or MIA with areas under the curve (AUCs) of 0.914, 0.956, and 0.833 for the training, tuning, and validation sets, respectively. The mean AUC of the radiologists was 0.835 in the validation set, without significant differences between radiologists and the model (p = 0.97). The sensitivity, specificity, and accuracy of the model were 71% (32/45), 87% (39/45), and 79% (71/90), respectively, whereas the corresponding values of the radiologists were 75.2% (203/270), 76.7% (207/270), and 75.9% (410/540) with a 5-mm threshold for the solid portion size.

CONCLUSIONS

The performance of the model for differentiating IPA from preinvasive lesions or MIA was comparable to that of the radiologists' measurements of solid portion size.

KEY POINTS

• A deep learning-based model differentiated IPA from preinvasive lesions or MIA with AUCs of 0.914 and 0.956 for the training and tuning sets, respectively. • In the validation set including subsolid nodules of 2 cm or smaller, the model showed an AUC of 0.833, being on par with the performance of the solid portion size measurements made by the radiologists (AUC, 0.835; p = 0.97). • SSNs with a solid portion measuring > 10 mm on CT showed a high probability of being IPA (positive predictive value, 93.5-100.0%).

Assessing invasiveness of subsolid lung adenocarcinomas with combined attenuation and geometric feature models

The aim of this study was to develop and test multiclass predictive models for assessing the invasiveness of individual lung adenocarcinomas presenting as subsolid nodules on computed tomography (CT). 227 lung adenocarcinomas were included: 31 atypical adenomatous hyperplasia and adenocarcinomas in situ (class H1), 64 minimally invasive adenocarcinomas (class H2) and 132 invasive adenocarcinomas (class H3). Nodules were segmented, and geometric and CT attenuation features including functional principal component analysis features (FPC1 and FPC2) were extracted. After a feature selection step, two predictive models were built with ordinal regression: Model 1 based on volume (log) (logarithm of the nodule volume) and FPC1, and Model 2 based on volume (log) and Q.875 (CT attenuation value at the 87.5% percentile). Using the 200-repeats Monte-Carlo cross-validation method, these models provided a multiclass classification of invasiveness with discriminative power AUCs of 0.83 to 0.87 and predicted the class probabilities with less than a 10% average error. The predictive modelling approach adopted in this paper provides a detailed insight on how the value of the main predictors contribute to the probability of nodule invasiveness and underlines the role of nodule CT attenuation features in the nodule invasiveness classification.

Determining the invasiveness of pure ground-glass nodules using dual-energy spectral computed tomography

Background

The present work aimed to investigate the clinical application of using quantitative parameters generated in the unenhanced phase (UP) and venous phase (VP) in dual-energy spectral CT for differentiating the invasiveness of pure ground-glass nodule (pGGN).

Methods

Sixty-two patients with 66 pGGNs who underwent preoperative dual-energy spectral CT in UP and VP were evaluated retrospectively. Nodules were divided into three groups based on pathology: adenocarcinoma in situ (AIS, n=19), minimally invasive adenocarcinoma (MIA, n=22) (both in the preinvasive lesion group) and invasive adenocarcinoma (IA, n=25). The iodine concentration (IC) and water content (WC) in nodules were measured in material decomposition images. The nodule CT numbers and slopes(k) were measured on monochromatic images. All measurements, including the maximum diameter of nodules were statistically compared between the AIS-MIA group and IA group.

Results

There were significant differences of WC in VP between AIS-MIA group and IA group (P<0.05). The CT attenuation values of the 40–140 keV monochromatic images in UP and VP were significantly higher for the invasive nodules. Logistic regression analysis showed that the maximum nodule diameter [odd ratio (OR) =1.21, 95% CI: 1.050–1.400, P<0.01] and CT number in 130 keV images in venous phase (OR =1.03, 95% CI: 1.014–1.047, P<0.001) independently predicted histological invasiveness.

Conclusions

The quantitative parameters in dual-energy spectral CT in the unenhanced phase and venous phase provide useful information in differentiating preinvasive lesion group from IA group of pGGN, especially the maximum nodule diameter and CT number in the 130 keV images in the venous phase.

A Subsolid Nodules Imaging Reporting System (SSN-IRS) for Classifying 3 Subtypes of Pulmonary Adenocarcinoma

OBJECTIVES

To develop an imaging reporting system for the classification of 3 adenocarcinoma subtypes of computed tomography (CT)-detected subsolid pulmonary nodules (SSNs) in clinical patients.

METHODS

Between November 2011 and October 2017, 437 pathologically confirmed SSNs were retrospectively identified. SSNs were randomly divided 2:1 into a training group (291 cases) and a testing group (146 cases). CT-imaging characteristics were analyzed using multinomial univariable and multivariable logistic regression analysis to identify discriminating factors for the 3 adenocarcinoma subtypes (pre-invasive lesions, minimally invasive adenocarcinoma, and invasive adenocarcinoma). These factors were used to develop a classification and regression tree model. Finally, an SSN Imaging Reporting System (SSN-IRS) was constructed based on the optimized classification model. For validation, the classification performance was evaluated in the testing group.

RESULTS

Of the CT-derived characteristics of SSNs, qualitative density (nonsolid or part-solid), core (non-core or core), semantic features (pleural indentation, vacuole sign, vascular invasion), and diameter of solid component (≤6 mm or >6 mm), were the most important factors for the SSN-IRS. The total sensitivity, specificity, and diagnostic accuracy of the SSN-IRS was 89.0% (95% confidence interval [CI], 84.8%-92.4%), 74.6% (95% CI, 70.8%-78.1%), and 79.4% (95% CI, 76.5%-82.0%) in the training group and 84.9% (95% CI, 78.1%-90.3%), 68.5% (95% CI, 62.8%-73.8%), and 74.0% (95% CI, 69.6%-78.0%) in the testing group, respectively.

CONCLUSIONS

The SSN-IRS can classify 3 adenocarcinoma subtypes using CT-based characteristics of subsolid pulmonary nodules. This classification tool can help clinicians to make follow-up recommendations or decisions for surgery in clinical patients with SSNs.

A radiomics model for determining the invasiveness of solitary pulmonary nodules that manifest as part-solid nodules

AIM

A nomogram model was developed to predict the histological subtypes of lung invasive adenocarcinomas (IAs) and minimally invasive adenocarcinomas (MIAs) that manifest as part-solid ground-glass nodules (GGNs).

MATERIALS AND METHODS

This retrospective study enrolled 119 patients with histopathologically confirmed part-solid GGNs assigned to the training (n=83) or testing cohorts (n=36). Radiomic features were extracted based on the unenhanced computed tomography (CT) images. R software was applied to process the qualitative and quantitative data. The CT features model, radiomic signature model, and combined prediction model were constructed and compared.

RESULTS

A total of 396 radiomic features were extracted from the preoperative CT images, four features including MaxIntensity, RMS, ZonePercentage, and LongRunEmphasis_angle0_offset7 were indicated to be the best discriminators to establish the radiomic signature model. The performance of the model was satisfactory in both the training and testing set with areas under the curve (AUCs) of 0.854 (95% confidence interval [CI]: 0.774 to 0.934) and 0.813 (95% CI: 0.670 to 0.955), respectively. The CT morphology of the lesion shape and diameter of the solid component were confirmed to be a significant feature for building the CT features model, which had an AUC of 0.755 (95% CI: 0.648 to 0.843). A nomogram that integrated lesion shape and radiomic signature was constructed, which contributed an AUC of 0.888 (95% CI: 0.82 to 0.955).

CONCLUSIONS

The radiomic signature could provide an important reference for differentiating IAs from MIAs, and could be significantly enhanced by the addition of CT morphology. The nomogram may be highly informative for making clinical decisions.

Histogram analysis combined with morphological characteristics to discriminate adenocarcinoma in situ or minimally invasive adenocarcinoma from invasive adenocarcinoma appearing as pure ground-glass nodule

OBJECTIVE

To construct a predictive model to discriminate adenocarcinoma in situ (AIS) or minimally invasive adenocarcinoma (MIA) from invasive adenocarcinoma (IAC) appearing as pure ground-glass nodules (pGGNs) using computed tomography (CT) histogram analysis combined with morphological characteristics and to evaluate its diagnostic performance.

MATERIALS AND METHODS

Two hundred eighty-nine patients with surgically resected solitary pGGN and pathologically diagnosed with AIS, MIA, or IAC in our institution from January 2014 to May 2018 were enrolled in our study. Two hundred twenty-six pGGNs (79 AIS, 84 MIA, and 63 IAC) were randomly selected and assigned to a model-development cohort, and the remaining 63 pGGNs (11 AIS, 29 MIA and 23 IAC) were assigned to a validation cohort. The morphological characteristics were established as model A and histogram parameters as model B. The diagnostic performances of model A, model B, and model A + B were evaluated and compared via receiver operating curve (ROC) analysis and logistic regression analysis.

RESULTS

Entropy (odd ratio [OR] = 23.25, 95%CI: 6.83-79.15, p < 0.001), microvascular sign (OR = 8.62, 95%CI: 3.72-19.98, p < 0.001) and the maximum diameter (OR = 4.37, 95%CI: 2.44-7.84, p < 0.001) were identified as independent predictors in the IAC group. The area under the ROC (Az value), accuracy, sensitivity and specificity of model A + B were 0.896, 88.1%, 79.4% and 91.4%, respectively, exhibiting a significantly higher Az value than either model A or model B alone (0.785 vs 0.896, p < 0.001; 0.849 vs 0.896, p = 0.029). Model A + B also conveyed a good diagnostic performance in the validation cohort, with an Az value of 0.851.

CONCLUSION

Histogram analysis combined with morphological characteristics exhibit a superior diagnostic performance in discriminating AIS-MIA from IAC appearing as pGGNs.

3D convolutional neural network for differentiating pre-invasive lesions from invasive adenocarcinomas appearing as ground-glass nodules with diameters ≤3 cm using HRCT

Background

Identification of pre-invasive lesions (PILs) and invasive adenocarcinomas (IACs) can facilitate treatment selection. This study aimed to develop an automatic classification framework based on a 3D convolutional neural network (CNN) to distinguish different types of lung cancer using computed tomography (CT) data.

Methods

The CT data of 1,545 patients suffering from pre-invasive or invasive lung cancer were collected from Fudan University Shanghai Cancer Center. All of the data were preprocessed through lung mask extraction and 3D reconstruction to adapt to different imaging scanners or protocols. The general flow for the classification framework consisted of nodule detection and cancer classification. The performance of our classification algorithm was evaluated using a receiver operating characteristic (ROC) analysis, with diagnostic results from three experienced radiologists.

Results

The sensitivity, specificity, accuracy, and AUC (area under the ROC curve) values of our proposed automatic classification method were 88.5%, 80.1%, 84.0%, and 89.2%, respectively. The results of the CNN classification method were compared to those of three experienced radiologists. The AUC value of our method (AUC =0.892) was higher than those of all radiologists (radiologist 1: 80.5%; radiologist 2: 83.9%; and radiologist 3: 86.7%).

Conclusions

The 3D CNN-based classification algorithm is a promising tool for the diagnosis of pre-invasive and invasive lung cancer and for the treatment choice decision.