Potential of artificial intelligence based on chest computed tomography to predict the nature of part-solid nodules

BACKGROUND

The potential of artificial intelligence (AI) to predict the nature of part-solid nodules based on chest computed tomography (CT) is still under exploration.

OBJECTIVE

To determine the potential of AI to predict the nature of part-solid nodules.

METHODS

Two hundred twenty-three patients diagnosed with part-solid nodules (241) by chest CT were retrospectively collected that were divided into benign group (104) and malignant group (137). Intraclass correlation coefficient (ICC) was used to assess the agreement in predicting malignancy, and the predictive effectiveness was compared between AI and senior radiologists. The parameters measured by AI and the size of solid components measured by senior radiologists were compared between two groups. Receiver operating characteristic (ROC) curve was chosen for calculating the Youden index of each quantitative parameter, which has statistical significance between two groups. Binary logistic regression performed on the significant indicators to suggest predictors of malignancy.

RESULTS

AI was in moderate agreement with senior radiologists (ICC = 0.686). The sensitivity, specificity and accuracy of two groups were close (p > 0.05). The longest diameter, volume and mean CT attenuation value and the largest diameter of solid components between benign and malignant groups were different significantly (p < 0.001). Logistic regression analysis showed that the longest diameter and mean CT attenuation value and the largest diameter of solid components were indicators for malignant part-solid nodules, the threshold of which were 9.45 mm, 425.0 HU and 3.45 mm, respectively.

CONCLUSION

Potential of quantitative parameter measured by AI to predict malignant part-solid nodules can provide a certain value for the clinical management.

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.

Tertiary lymphoid structures in lung adenocarcinoma: characteristics and related factors

OBJECTIVE

Tertiary lymphoid structures (TLSs) are found in a variety of malignancies and affect the growth of tumors, but few studies have addressed their role in lung adenocarcinoma (LAC). We aimed to evaluate clinical features associated with TLSs in patients with LAC.

METHODS AND MATERIALS

A collection of resected pulmonary nodules in patients with LAC was retrospectively analyzed. TLSs were quantified by their number per square millimeter tumor area (density) and by the degree of lymphocyte aggregation (maturity) in each case. The correlation between TLS density and maturity and clinical features was calculated.

RESULTS

A total of 243 patients were selected, of whom 219 exhibited TLSs. The occurrence of TLSs was correlated with computed tomography (CT) features as follows: pure ground-glass nodules (pGGNs) (n = 43) was associated with a lower occurrence rate than part-solid nodules (PSNs) (n = 112) and solid nodules (SNs) were (n = 88) (p = 0.037). TLS density was correlated with age and CT features. Poisson regression showed higher TLS density in PSNs and SNs than in pGGNs (incidence rate ratio [IRR]: 3.137; 95% confidence interval [CI]: 1.35-7.27; p = 0.008 and IRR: 2.44; 95% CI: 1.02-5.85; p = 0.046, respectively). In addition, TLS density was higher in patients aged under 60 years than in those aged over 60 years (IRR: 0.605; 95% CI: 0.4-0.92; p = 0.018). The maturity of TLSs was higher in patients with higher tumor stages (p = 0.026).

CONCLUSIONS

We demonstrated distinct profiles of TLSs in early LAC and their correlations with CT features, age, and tumor stages, which could help understand tumor progression and management.

Evaluation of solid portions in non-small cell lung cancer-the solid part is not always measurable for clinical T factor

BACKGROUND

Solid component size on thin-section computed tomography is used for T-staging according to the eighth edition of the Tumor Node Metastasis classification of lung cancer. However, the feasibility of using the solid component to measure clinical T-factor remains controversial.

METHODS

We evaluated the feasibility of measuring the solid component in 859 tumours, which were suspected cases of primary lung cancers, requiring surgical resection regardless of the procedure or clinical stage. After excluding 126 pure ground-glass opacity tumours and 450 solid tumours, 283 part-solid tumours were analysed to determine the frequency of cases where the measurement of the solid portion was difficult along with the associated cause. Pathological invasiveness was also evaluated.

RESULTS

The solid portion of 10 lesions in 283 part-solid nodules was difficult to measure due to an underlying lung disease (emphysema and pneumonitis). The solid portion of 62 lesions (21.9%) without emphysema and pneumonitis was difficult to measure due to imaging features of the tumours. Among the 62 patients, five had no malignancy and one with a tumour size of 33 mm had nodal metastasis. There were 56 lesions with a tumour size of ≤30 mm, wherein nodal metastases, vascular and/or lymphatic invasions were not observed.

CONCLUSION

For one-fifth of the part-solid tumours, measurement of the solid component was difficult. Moreover, these lesions had low invasiveness, especially in T1. The measurement of the solid portion and the classification of T1 in 1-cm increments may be complex.

Distinct clinicopathologic factors and prognosis based on the presence of ground-glass opacity components in patients with resected stage I non-small cell lung cancer

Background

This study was to investigate the prognostic value of ground-glass opacity(GGO) components and to evaluate distinct the clinicopathological variables of survival outcomes for the pure-GGO, part-solid and solid groups of patients with resected stage I non-small cell lung cancer (NSCLC).

Methods

We retrospectively reviewed the structured data for stage I NSCLC patients who had undergone the curative-intent surgical resection in the Lung Cancer Database of West China Hospital from 2009 to 2016. The eligible patients were divided into the pure-GGO, part-solid and solid groups according to the radiological manifestation. Univariate and multivariate Cox regression analyses were performed between the 3 groups. And we further evaluated the clinicopathological variables in each group separately.

Results

Among a total of 2,775 eligible patients enrolled into the cohort were 1,587 (57.19%) in the solid group, 508 (18.31%) in the part-solid group, and 680 (24.50%) in the pure-GGO group. The 5-year overall survival (OS) and recurrence-free survival (RFS) rates were 98.8% and 98.0% in the pure-GGO group, 96.0% and 86.5% in the part-solid group, and 88.0% and 75.5% in the solid group, respectively (P<0.001). Presence of GGO components was a significantly favorable prognosticator (HR =0.415, 95% CI: 0.286–0.601). Different groups had distinct prognostic factors. LVI was the shared risk factor for groups with presence of GGO components in both part-solid and pure-GGO groups. Pathological stage (IA or IB) was influential exclusively for the pure-GGO group. In the solid group, females, younger patients, and patients without VPI had better survival. But such independent significance did not exist in the other two groups.

Conclusions

GGO component was a strong prognosticator of better prognosis in resected patients with stage I NSCLC. Prognostic factors and survival outcomes were disparate among the pure-GGO, part-solid, and solid group. Our results support the proposal that the next edition tumor-node-metastasis (TNM) classification should consider the importance of GGO components as a new T descriptor.

The prognostic value of Kirsten rat sarcoma viral oncogene homolog mutations in resected lung adenocarcinoma differs according to clinical features

BACKGROUND

The ninth edition of lung cancer staging system recommends that specific driver mutations should be considered as prognostic factors in survival models. This study comprehensively investigated the prognostic value of Kirsten rat sarcoma viral oncogene homolog (KRAS) mutation in patients with resected lung adenocarcinomas according to different clinicopathologic and radiologic characteristics.

METHODS

In total, 1464 patients with completely resected primary lung adenocarcinomas were examined for KRAS mutations from November 2008 to March 2015. Age, sex, smoking status, performance status, tumor-node-metastasis stage, radiologic features, and histologic subtypes were collected. Competing risk model was used to estimate the cumulative incidence rate of recurrence. Cox regression multivariable analyses on recurrence-free survival (RFS) and overall survival (OS) were performed.

RESULTS

KRAS mutations were more frequent in male subjects (P < .001), current/former smokers (P < .001), invasive mucinous adenocarcinoma (P < .001), and solid tumors (P < .001). In general, KRAS-mutated patients had greater cumulative recurrence rate (hazard ratio [HR], 1.95; 95% confidence interval [CI], 1.23-3.08; P < .001) and worse overall survival (OS; HR, 1.88; 95% CI, 1.23-2.87; P < .001) than KRAS wild-type patients. The OS (P < .001) of patients harboring KRAS-G12C/V mutations was shorter than that of other KRAS-mutated patients. Cox multivariable analyses demonstrated that KRAS mutations were independently associated with worse RFS (HR, 5.34; 95% CI, 2.53-11.89; P = .001) and OS (HR, 2.63; 95% CI, 1.03-6.76; P = .044) in part-solid lung adenocarcinomas. For stage I patients, Cox multivariable analyses revealed that KRAS mutation was an independent risk factor for RFS (HR, 2.05; 95% CI, 1.19-3.56; P = .010) and OS (HR, 2.38; 95% CI, 1.29-4.40; P = .005).

CONCLUSIONS

In this study, we revealed that KRAS mutations was an independent prognostic factor in part-solid tumors and in stage I lung adenocarcinomas. These findings may contribute to the ninth edition of lung cancer staging project.

What's New on Quantitative CT Analysis as a Tool to Predict Growth in Persistent Pulmonary Subsolid Nodules? A Literature Review

Pulmonary subsolid nodules (SSNs) are observed not infrequently on thin-section chest computed tomography (CT) images. SSNs persisting after a follow-up period of three to six months have a high likelihood of being pre-malignant or malignant lesions. Malignant SSNs usually represent the histologic spectrum of pulmonary adenocarcinomas, and pulmonary adenocarcinomas presenting as SSNs exhibit quite heterogeneous behavior. In fact, while most lesions show an indolent course and may grow very slowly or remain stable for many years, others may exhibit significant growth in a relatively short time. Therefore, it is not yet clear which persistent SSNs should be surgically removed and for how many years stable SSNs should be monitored. In order to solve these two open issues, the use of quantitative analysis has been proposed to define the "tailored" management of persistent SSNs. The main purpose of this review was to summarize recent results about quantitative CT analysis as a diagnostic tool for predicting the behavior of persistent SSNs. Thus, a literature search was conducted in PubMed/MEDLINE, Scopus, and Web of Science databases to find original articles published from January 2014 to October 2019. The results of the selected studies are presented and compared in a narrative way.