The imaging of small pulmonary nodules

Convolutional Neural Network-Based Diagnostic Model for a Solid, Indeterminate Solitary Pulmonary Nodule or Mass on Computed Tomography

PURPOSE

To establish a non-invasive diagnostic model based on convolutional neural networks (CNNs) to distinguish benign from malignant lesions manifesting as a solid, indeterminate solitary pulmonary nodule (SPN) or mass (SPM) on computed tomography (CT).

METHOD

A total of 459 patients with solid indeterminate SPNs/SPMs on CT were ultimately included in this retrospective study and assigned to the train (n=366), validation (n=46), and test (n=47) sets. Histopathologic analysis was available for each patient. An end-to-end CNN model was proposed to predict the natural history of solid indeterminate SPN/SPMs on CT. Receiver operating characteristic curves were plotted to evaluate the predictive performance of the proposed CNN model. The accuracy, sensitivity, and specificity of diagnoses by radiologists alone were compared with those of diagnoses by radiologists by using the CNN model to assess its clinical utility.

RESULTS

For the CNN model, the AUC was 91% (95% confidence interval [CI]: 0.83-0.99) in the test set. The diagnostic accuracy of radiologists with the CNN model was significantly higher than that without the model (89 vs. 66%, P<0.01; 87 vs. 61%, P<0.01; 85 vs. 66%, P=0.03, in the train, validation, and test sets, respectively). In addition, while there was a slight increase in sensitivity, the specificity improved significantly by an average of 42% (the corresponding improvements in the three sets ranged from 43, 33, and 42% to 82, 78, and 84%, respectively; P<0.01 for all).

CONCLUSION

The CNN model could be a valuable tool in non-invasively differentiating benign from malignant lesions manifesting as solid, indeterminate SPNs/SPMs on CT.

Epithelioid hemangioendothelioma: evaluation by 18F-FDG PET/CT

PURPOSE

The purpose of this study was to evaluate the imaging characteristics of epithelioid hemangioendothelioma (EHE) on staging ¹⁸F-FDG PET/CT.

MATERIALS AND METHODS

An IRB-approved retrospective review was conducted for patients with biopsy-proven EHE who underwent FDG PET/CT at our institution between 2005 and 2019. Patients with a history of surgery, chemotherapy, or radiotherapy prior to PET/CT were excluded. PET/CT exams were analyzed, noting metabolic activity, distribution of involvement, and CT morphologic features. PET/CT findings were correlated with comparative CT and MRI performed within three months.

RESULTS

There were 35 patients [21 females, 14 males; average age 55.1±16.9 years (range 15-82 years)]. 18/35 patients (52%) had more than one organ affected on PET/CT. The most common sites were liver [21/35 (60%)], lung [(19/35 (54%)], bone [5/35 (14%)], lymph nodes [4/35 (11%)], and vasculature [4/35 (11%)]. Most patients [30/35, (86%)] presented with multiple lesions. The average largest lesion dimension was 4.0±3.6 cm (range 0.6-15.0 cm). The average SUVmax of the most metabolically active lesion at any site was 5.3±3.3 (range 1.2-17.1), and for bone was 7.9±5.4 (range 3.5-17.1), liver was 5.1±2.1 (range 2.6-10.5), and lung was 3.0±1.9 (range 1.2-8.5). Of patients with pulmonary lesions, 9/19 (47%) showed calcification, and 4/19 (21%) had nodules that were either non FDG-avid or too small for accurate SUV assessment. Of patients with hepatic lesions, 11/21 (52%) demonstrated capsular retraction, and 12/21 (57%) were found to have additional hepatic lesions on contrast-enhanced CT or MRI that were occult on PET/CT.

CONCLUSION

EHE demonstrates variable, but most commonly moderate FDG activity on PET/CT. The most common sites of disease are the liver, lungs, and bones, and most patients present with multiple lesions and more than one organ involved. Given the intrinsic metabolic activity and multi-organ involvement, FDG PET/CT represents an attractive modality for EHE evaluation. However, it may be best used in combination with CT or MRI given that EHE pulmonary or hepatic lesions may be missed by PET/CT.

Differentiating Peripherally-Located Small Cell Lung Cancer From Non-small Cell Lung Cancer Using a CT Radiomic Approach

Lung cancer can be classified into two main categories: small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), which are different in treatment strategy and survival probability. The lung CT images of SCLC and NSCLC are similar such that their subtle differences are hardly visually discernible by the human eye through conventional imaging evaluation. We hypothesize that SCLC/NSCLC differentiation could be achieved via computerized image feature analysis and classification in feature space, as termed a radiomic model. The purpose of this study was to use CT radiomics to differentiate SCLC from NSCLC adenocarcinoma. Patients with primary lung cancer, either SCLC or NSCLC adenocarcinoma, were retrospectively identified. The post-diagnosis pre-treatment lung CT images were used to segment the lung cancers. Radiomic features were extracted from histogram-based statistics, textural analysis of tumor images and their wavelet transforms. A minimal-redundancy-maximal-relevance method was used for feature selection. The predictive model was constructed with a multilayer artificial neural network. The performance of the SCLC/NSCLC adenocarcinoma classifier was evaluated by the area under the receiver operating characteristic curve (AUC). Our study cohort consisted of 69 primary lung cancer patients with SCLC (n = 35; age mean ± SD = 66.91± 9.75 years), and NSCLC adenocarcinoma (n = 34; age mean ± SD = 58.55 ± 11.94 years). The SCLC group had more male patients and smokers than the NSCLC group (P < 0.05). Our SCLC/NSCLC classifier achieved an overall performance of AUC of 0.93 (95% confidence interval = [0.85, 0.97]), sensitivity = 0.85, and specificity = 0.85). Adding clinical data such as smoking history could improve the performance slightly. The top ranking radiomic features were mostly textural features. Our results showed that CT radiomics could quantitatively represent tumor heterogeneity and therefore could be used to differentiate primary lung cancer subtypes with satisfying results. CT image processing with the wavelet transformation technique enhanced the radiomic features for SCLC/NSCLC classification. Our pilot study should motivate further investigation of radiomics as a non-invasive approach for early diagnosis and treatment of lung cancer.

The multidisciplinary team plays an important role in the prediction of small solitary pulmonary nodules: a propensity-score-matching study

Background

According to guidelines, it is recommended that pulmonary nodules be discussed by a multidisciplinary team (MDT); however, the evidence for the effectiveness of MDT is sparse. To demonstrate the importance of the involvement of an MDT for the prediction of small solitary pulmonary nodules, we conducted this retrospective study.

Methods

The patient database of those who attended our MDT and the electronic medical record system of our hospital was used; we collected all the data from patients found with small solitary pulmonary nodules (≤2 cm), which were suspected as malignant and who received a resection of the nodules. We summarized their characteristics and analyzed them, and then compared the post-operation pathological diagnosis of the patients who attended an MDT to those who did not participate in an MDT during the same period (2017-2019.2). We also collected the follow-up data. Propensity-score-matching was utilized during the process of analysis to get a more reliable conclusion.

Results

Most of the qualified patients were female. Most of the small solitary pulmonary nodules (≤2 cm) were adenocarcinoma and located on the right upper lobe. There were no differences in the SUV value between malignant nodules and benign nodules. After propensity-score matching, the total positive prediction value of small solitary pulmonary nodules (≤2 cm) without an MDT was 69.4%, while that with MDT was 77.6%; the difference was not significant with a P value of 0.30. The negative predictive value of MDT was 76.2%.

Conclusions

In developing countries, small solitary pulmonary nodules tend to be more correctly diagnosed with MDT.

Predicting malignant potential of subsolid nodules: can radiomics preempt longitudinal follow up CT?

Background

To assess if radiomics can differentiate benign and malignant subsolid lung nodules (SSNs) on baseline or follow up chest CT examinations. If radiomics can differentiate between benign and malignant subsolid lung nodules, the clinical implications are shorter follow up CT imaging and early recognition of lung adenocarcinoma on imaging.

Materials and methods

The IRB approved retrospective study included 36 patients (mean age 69 ± 8 years; 5 males, 31 females) with 108 SSNs (31benign, 77 malignant) who underwent follow up chest CT for evaluation of indeterminate SSN. All SSNs were identified on both baseline and follow up chest CT. DICOM CT images were deidentified and exported into the open access 3D Slicer software (version 4.7) to obtain radiomic features. Logistic regression analyses and receiver operating characteristic (ROC) curves for various quantitative parameters were generated with SPSS statistical software.

Results

Only 2/92 radiomic features (cluster shade and surface volume ratio) enabled differentiation between malignant and benign SSN on baseline chest CT (P = 0.01 and 0.03) with moderate accuracy [AUC 0.624 (0.505–0.743)]. On follow-up CT, 52/92 radiomic features were significantly different between benign and malignant SSN (P: 0.04 - < 0.0001) with improved accuracy [AUC: 0.708 (0.605–0.811), P = 0.04 - < 0.0001]. Radiomics of benign SSN were stable over time, whereas 63/92 radiomic features of malignant SSNs changed significantly between the baseline and follow up chest CT (P: 0.04 - < 0.0001).

Conclusions

Temporal changes in radiomic features of subsolid lung nodules favor malignant etiology over benign. The change in radiomics features of subsolid lung nodules can allow shorter follow up CT imaging and early recognition of lung adenocarcinoma on imaging. Radiomic features have limited application in differentiating benign and early malignant SSN on baseline chest CT.

Small pulmonary granuloma is often misdiagnosed as lung cancer by positron emission tomography/computer tomography in diabetic patients

OBJECTIVES

A small pulmonary granuloma (SPG) is often misdiagnosed as lung cancer in diabetic patients by positron emission tomography/computed tomography (PET/CT). The present study was conducted to investigate whether diabetes is the influencing factor and to determine other related factors that have an impact on the diagnostic results following PET/CT examination.

METHODS

All clinical, imaging and pathological data of patients diagnosed with pulmonary nodules by PET/CT from January 2004 to December 2017 in our department were collected. Patients with an SPG who were wrongly diagnosed with lung cancer by PET/CT were enrolled (n = 79). The propensity score matching method was used to create a comparable control adenocarcinoma group (n = 395). Maximum standard uptake values, diabetes and fasting blood-glucose (FBG) were determined and analysed.

RESULTS

The average maximum standard uptake values in the 2 groups were comparable (P = 0.801). Maximum standard uptake values in 5 subsections were not significantly different between the 2 groups (P = 0.135). The odds ratio (OR) of 3.326 [95% confidence interval (CI) 1.671-6.623] for diabetes favoured misdiagnosis and was statistically significant (P < 0.001). Furthermore, in patients with high FBG levels (≥7.0 mmol/l), the risk of misdiagnosis of SPG increased significantly compared with normal FBG level (OR 2.601, 95% CI 1.174-5.761; P = 0.015).

CONCLUSIONS

Diabetes and high FBG level were the influencing factors in the false-positive results of lung cancer by PET/CT examination.

Correlation of quantified metabolic activity in nonsmall cell lung cancer with tumor size and tumor pathological characteristics

The aim of this study was to evaluate the relationship between maximum standardized uptake value (SUVmax) with tumor size and tumor pathological characteristics as well as suggesting equations between SUVmax and tumor size in patients with nonsmall cell lung cancer (NSCLC) to help differentiate between pathology types.We retrospectively analyzed the fluorodeoxyglucose-positron emission tomography/computed tomography (FDG-PET/CT) findings of 98 patients with NSCLC. Statistical differences were considered significant when P < .05. Correlation between SUVmax and other variables was determined by Pearson and Spearman correlation. Both linear and nonlinear regression analysis were used to determine equations between SUVmax and tumor size to help differentiate between pathology types.The mean SUVmax in patients with squamous cell carcinoma was significantly higher than that of adenocarcinoma (21.35 ± 1.73 vs 13.75 ± 0.89, P = .000). The results of regression analysis indicated that among all equations determined with relative accuracy, the "cubic equation" has the highest accuracy when considering the relationship between SUVmax and tumor size in patients with adenocarcinoma. In patients with squamous cell carcinoma, the most accurate equation was obtained using the "quadratic equation."There was a significant correlation between SUVmax and tumor differentiation and tumor size in patients with adenocarcinoma. SUVmax of patients with squamous cell carcinoma also had a significant correlation with tumor size. Overall SUVmax of patients with NSCLC could be predicted by tumor size value. In patients with squamous cell carcinoma compared with those with adenocarcinoma, SUVmax with less accuracy can be determined by tumor size. Linear regression analysis line slope can be used as an index for distinguishing adenocarcinoma from squamous cell carcinoma.