Reinforcement learning for individualized lung cancer screening schedules: A nested case-control study.

BACKGROUND: The current guidelines for managing screen-detected pulmonary nodules offer rule-based recommendations for immediate diagnostic work-up or follow-up at intervals of 3, 6, or 12 months. Customized visit plans are lacking. PURPOSE: To develop individualized screening schedules using reinforcement learning (RL) and evaluate the effectiveness of RL-based policy models. METHODS: Using a nested case-control design, we retrospectively identified 308 patients with cancer who had positive screening results in at least two screening rounds in the National Lung Screening Trial. We established a control group that included cancer-free patients with nodules, matched (1:1) according to the year of cancer diagnosis. By generating 10,164 sequence decision episodes, we trained RL-based policy models, incorporating nodule diameter alone, combined with nodule appearance (attenuation and margin) and/or patient information (age, sex, smoking status, pack-years, and family history). We calculated rates of misdiagnosis, missed diagnosis, and delayed diagnosis, and compared the performance of RL-based policy models with rule-based follow-up protocols (National Comprehensive Cancer Network guideline; China Guideline for the Screening and Early Detection of Lung Cancer). RESULTS: We identified significant interactions between certain variables (e.g., nodule shape and patient smoking pack-years, beyond those considered in guideline protocols) and the selection of follow-up testing intervals, thereby impacting the quality of the decision sequence. In validation, one RL-based policy model achieved rates of 12.3% for misdiagnosis, 9.7% for missed diagnosis, and 11.7% for delayed diagnosis. Compared with the two rule-based protocols, the three best-performing RL-based policy models consistently demonstrated optimal performance for specific patient subgroups based on disease characteristics (benign or malignant), nodule phenotypes (size, shape, and attenuation), and individual attributes. CONCLUSIONS: This study highlights the potential of using an RL-based approach that is both clinically interpretable and performance-robust to develop personalized lung cancer screening schedules. Our findings present opportunities for enhancing the current cancer screening system.

Acceptability of adding a non-contrast abdominal CT scan to screen for kidney cancer and other abdominal pathology within a community-based CT screening programme for lung cancer: A qualitative study.

OBJECTIVES: The Yorkshire Kidney Screening Trial (YKST) is a feasibility study of adding non-contrast abdominal CT scanning to screen for kidney cancer and other abdominal malignancies to community-based CT screening for lung cancer within the Yorkshire Lung Screening Trial (YLST). This study explored the acceptability of the combined screening approach to participants and healthcare professionals (HCPs) involved in the trial. METHODS: We conducted semi-structured interviews with eight HCPs and 25 participants returning for the second round of scanning within YLST, 20 who had taken up the offer of the additional abdominal CT scan and five who had declined. Transcripts were analysed using thematic analysis, guided by the Theoretical Framework of Acceptability. RESULTS: Overall, combining the offer of a non-contrast abdominal CT scan alongside the low-dose thoracic CT was considered acceptable to participants, including those who had declined the abdominal scan. The offer of the additional scan made sense and fitted well within the process, and participants could see benefits in terms of efficiency, cost and convenience both for themselves as individuals and also more widely for the NHS. Almost all participants made an instant decision at the point of initial invitation based more on trust and emotions than the information provided. Despite this, there was a clear desire for more time to decide whether to accept the scan or not. HCPs also raised concerns about the burden on the study team and wider healthcare system arising from additional workload both within the screening process and downstream following findings on the abdominal CT scan. CONCLUSIONS: Adding a non-contrast abdominal CT scan to community-based CT screening for lung cancer is acceptable to both participants and healthcare professionals. Giving potential participants prior notice and having clear pathways for downstream management of findings will be important if it is to be offered more widely.

Intrathoracic Lymph Node Microcalcifications are Associated With a High Prevalence of Malignancy and Anaplastic Lymphoma Kinase Rearrangement: The "Calce" Study.

BACKGROUND: Microcalcifications are acknowledged as a malignancy risk factor in multiple cancers. However, the prevalence and association of intrathoracic lymph node (ILN) calcifications with malignancy remain unexplored. METHODS: In this cross-sectional study, we enrolled patients with known/suspected malignancy and an indication for endosonography for diagnosis or ILN staging. We assessed the prevalence and pattern of calcified ILNs and the prevalence of malignancy in ILNs with and without calcifications. In addition, we evaluated the genomic profile and PD-L1 expression in lung cancer patients, stratifying them based on the presence or absence of ILN calcifications. RESULTS: A total of 571 ILNs were sampled in 352 patients. Calcifications were detected in 85 (24.1%) patients and in 94 (16.5%) ILNs, with microcalcifications (78/94, 83%) being the predominant type. Compared with ILNs without calcifications (214/477, 44.9%), the prevalence of malignancy was higher in ILNs with microcalcifications (73/78, 93.6%; P<0.0001) but not in those with macrocalcifications (7/16, 43.7%; P=0.93). In patients with lung cancer, the high prevalence of metastatic involvement in ILNs displaying microcalcifications was independent of lymph node size (< or >1 cm) and the clinical stage (advanced disease; cN2/N3 disease; cN0/N1 disease). The anaplastic lymphoma kinase (ALK) rearrangement was significantly more prevalent in patients with than in those without calcified ILNs (17.4% vs. 1.7%, P<0.001), and all of them exhibited microcalcifications. CONCLUSION: ILN microcalcifications are common in patients undergoing endosonography for suspected malignancy, and they are associated with a high prevalence of metastatic involvement and ALK rearrangement.

3D airway geometry analysis of factors in airway navigation failure for lung nodules.

BACKGROUND: This study aimed to quantitatively reveal contributing factors to airway navigation failure during radial probe endobronchial ultrasound (R-EBUS) by using geometric analysis in a three-dimensional (3D) space and to investigate the clinical feasibility of prediction models for airway navigation failure. METHODS: We retrospectively reviewed patients who underwent R-EBUS between January 2017 and December 2018. Geometric quantification was analyzed using in-house software built with open-source python libraries including the Vascular Modeling Toolkit ( http://www.vmtk.org ), simple insight toolkit ( https://sitk.org ), and sci-kit image ( https://scikit-image.org ). We used a machine learning-based approach to explore the utility of these significant factors. RESULTS: Of the 491 patients who were eligible for analysis (mean age, 65 years +/- 11 [standard deviation]; 274 men), the target lesion was reached in 434 and was not reached in 57. Twenty-seven patients in the failure group were matched with 27 patients in the success group based on propensity scores. Bifurcation angle at the target branch, the least diameter of the last section, and the curvature of the last section are the most significant and stable factors for airway navigation failure. The support vector machine can predict airway navigation failure with an average area under the curve of 0.803. CONCLUSIONS: Geometric analysis in 3D space revealed that a large bifurcation angle and a narrow and tortuous structure of the closest bronchus from the lesion are associated with airway navigation failure during R-EBUS. The models developed using quantitative computer tomography scan imaging show the potential to predict airway navigation failure.

Artificial intelligence's contribution to early pulmonary lesion detection in chest X-rays: insights from two retrospective studies on a Czech population.

In recent years healthcare is undergoing significant changes due to technological innovations, with Artificial Intelligence (AI) being a key trend. Particularly in radiodiagnostics, according to studies, AI has the potential to enhance accuracy and efficiency. We focus on AI's role in diagnosing pulmonary lesions, which could indicate lung cancer, based on chest X-rays. Despite lower sensitivity in comparison to other methods like chest CT, due to its routine use, X-rays often provide the first detection of lung lesions. We present our deep learning-based solution aimed at improving lung lesion detection, especially during early-stage of illness. We then share results from our previous studies validating this model in two different clinical settings: a general hospital with low prevalence findings and a specialized oncology center. Based on a quantitative comparison with the conclusions of radiologists of different levels of experience, our model achieves high sensitivity, but lower specificity than comparing radiologists. In the context of clinical requirements and AI-assisted diagnostics, the experience and clinical reasoning of the doctor play a crucial role, therefore we currently lean more towards models with higher sensitivity over specificity. Even unlikely suspicions are presented to the doctor. Based on these results, it can be expected that in the future artificial intelligence will play a key role in the field of radiology as a supporting tool for evaluating specialists. To achieve this, it is necessary to solve not only technical but also medical and regulatory aspects. It is crucial to have access to quality and reliable information not only about the benefits but also about the limitations of machine learning and AI in medicine.

Diagnostic efficiency of intravoxel incoherent motion-based virtual magnetic resonance elastography in pulmonary neoplasms.

BACKGROUND: The aim of the study were as below. (1) To investigate the feasibility of intravoxel incoherent motion (IVIM)-based virtual magnetic resonance elastography (vMRE) to provide quantitative estimates of tissue stiffness in pulmonary neoplasms. (2) To verify the diagnostic performance of shifted apparent diffusion coefficient (sADC) and reconstructed virtual stiffness values in distinguishing neoplasm nature. METHODS: This study enrolled 59 patients (37 males, 22 females) with one pulmonary neoplasm who underwent computed tomography-guided percutaneous transthoracic needle biopsy (PTNB) with pathological diagnosis (26 adenocarcinoma, 10 squamous cell carcinoma, 3 small cell carcinoma, 4 tuberculosis and 16 non-specific benign; mean age, 60.81 ± 9.80 years). IVIM was performed on a 3 T magnetic resonance imaging scanner before biopsy. sADC and virtual shear stiffness maps reflecting lesion stiffness were reconstructed. sADC and virtual stiffness values of neoplasm were extracted, and the diagnostic performance of vMRE in distinguishing benign and malignant and detailed pathological type were explored. RESULTS: Compared to benign neoplasms, malignant ones had a significantly lower sADC and a higher virtual stiffness value (P < 0.001). Subsequent subtype analyses showed that the sADC values of adenocarcinoma and squamous cell carcinoma groups were significantly lower than non-specific benign group (P = 0.013 and 0.001, respectively). Additionally, virtual stiffness values of the adenocarcinoma and squamous cell carcinoma subtypes were significantly higher than non-specific benign group (P = 0.008 and 0.001, respectively). However, no significant correlation was found among other subtype groups. CONCLUSIONS: Non-invasive vMRE demonstrated diagnostic efficiency in differentiating the nature of pulmonary neoplasm. vMRE is promising as a new method for clinical diagnosis.

Placental Transmogrification of the Lung Presenting as Cystic-Solid Mass: A Case Report.

The gradually progressive solitary cystic-solid mass of chest CT scans is highly suggestive of lung cancer. We report a case of a 29-year-old woman with a persistent cystic-solid lesion in the right upper lobe. A chest CT scan showed a 35 mm × 44 mm × 51 mm focal cystic-solid mass in the anterior segment of the right upper lobe. The size of lesion had increased over 3 years, especially for the solid component. The right upper lobe pneumonectomy was performed. Postoperative pathological examination showed placental transmogrification of the lung, which is a rare cause of pulmonary cystic lesion.

Enhanced deep learning model for precise nodule localization and recurrence risk prediction following curative-intent surgery for lung cancer.

PURPOSE: Radical surgery is the primary treatment for early-stage resectable lung cancer, yet recurrence after curative surgery is not uncommon. Identifying patients at high risk of recurrence using preoperative computed tomography (CT) images could enable more aggressive surgical approaches, shorter surveillance intervals, and intensified adjuvant treatments. This study aims to analyze lung cancer sites in CT images to predict potential recurrences in high-risk individuals. METHODS: We retrieved anonymized imaging and clinical data from an institutional database, focusing on patients who underwent curative pulmonary resections for non-small cell lung cancers. Our study used a deep learning model, the Mask Region-based Convolutional Neural Network (MRCNN), to predict cancer locations and assign recurrence classification scores. To find optimized trained weighted values in the model, we developed preprocessing python codes, adjusted dynamic learning rate, and modifying hyper parameter in the model. RESULTS: The model training completed; we performed classifications using the validation dataset. The results, including the confusion matrix, demonstrated performance metrics: bounding box (0.390), classification (0.034), mask (0.266), Region Proposal Network (RPN) bounding box (0.341), and RPN classification (0.054). The model successfully identified lung cancer recurrence sites, which were then accurately mapped onto chest CT images to highlight areas of primary concern. CONCLUSION: The trained model allows clinicians to focus on lung regions where cancer recurrence is more likely, acting as a significant aid in the detection and diagnosis of lung cancer. Serving as a clinical decision support system, it offers substantial support in managing lung cancer patients.

ALINA: Another stepping stone toward a future of LDCT lung cancer screening in never-smokers?

The ALINA trial1 demonstrated that 2 years of adjuvant alectinib achieved statistically significantly improved 2-year overall and central nervous system (CNS) disease-free survival over platinum-doublet chemotherapy in resected early-stage (IB ≥ 4 cm to IIIA) ALK+ non-small cell lung cancer (NSCLC). Identifying early-stage ALK+ NSCLC patients (60% were never-smokers in the ALINA trial) may require low-dose computed tomography (LDCT) lung cancer screening in never-smokers.

The Value of Radiomics Features Based on HRCT in Predicting whether the Lung Sub-Centimeter Pure Ground Glass Nodule is Benign or Malignant.

OBJECTIVE: In this study, a radiomics model was created based on High-Resolution Computed Tomography (HRCT) images to noninvasively predict whether the sub-centimeter pure Ground Glass Nodule (pGGN) is benign or malignant. METHODS: A total of 235 patients (251 sub-centimeter pGGNs) who underwent preoperative HRCT scans and had postoperative pathology results were retrospectively evaluated. The nodules were randomized in a 7:3 ratio to the training (n=175) and the validation cohort (n=76). The volume of interest was delineated in the thin-slice lung window, from which 1316 radiomics features were extracted. The Least Absolute Shrinkage and Selection Operator (LASSO) was used to select the radiomics features. Univariate and multivariable logistic regression were used to evaluate the independent risk variables. The performance was assessed by obtaining Receiver Operating Characteristic (ROC) curves for the clinical, radiomics, and combined models, and then the Decision Curve Analysis (DCA) assessed the clinical applicability of each model. RESULTS: Sex, volume, shape, and intensity mean were chosen by univariate analysis to establish the clinical model. Two radiomics features were retained by LASSO regression to build the radiomics model. In the training cohort, the Area Under the Curve (AUC) of the radiomics (AUC=0.844) and combined model (AUC=0.871) was higher than the clinical model (AUC=0.773). In evaluating whether or not the sub-centimeter pGGN is benign, the DCA demonstrated that the radiomics and combined model had a greater overall net benefit than the clinical model. CONCLUSION: The radiomics model may be useful in predicting the benign and malignant sub-centimeter pGGN before surgery.

Patient satisfaction with letter-based communication of LCS pulmonary nodule results.

OBJECTIVE: To analyze patient satisfaction with letter-based communication of lung cancer screening (LCS) pulmonary nodule results. STUDY DESIGN: Prospective randomized controlled trial of LCS between May and December 2019. METHODS: All participants came from a prospective randomized controlled study on pulmonary nodule results in LCS with low-dose CT (LDCT) to analyze patient satisfaction, perception of information received via letters, preferred methods of receiving results, and dissatisfaction-related characteristics. RESULTS: A total of 153 patients were detected to have pulmonary nodules among 600 recruited participants in the lung cancer high-risk group screened using LDCT. Most of the patients were satisfied with receiving pulmonary nodule results via letters (78.4%; n = 120) and agreed that the letters contained an appropriate amount of information (83.7%; n = 128). Univariate logistic regression analysis revealed that satisfaction was related to age (OR, 0.905; 95% CI, 0.832-0.985), education level (OR, 0.367; 95% CI, 0.041-3.250), no family history of cancer (OR, 0.100; 95% CI, 0.011-0.914), and the number of nodules (OR, 6.028; 95% CI, 1.641-22.141). Of the patients who reported dissatisfaction with letter-based communication (7.2%; n = 11), the most common reasons cited were that they contained insufficient patient education materials and that it was difficult to comprehend the medical terminology. The majority of participants (61.4%; n = 94) reported that they would prefer the letter-based communication. No correlation was identified between satisfaction and gender, smoking status, alcohol consumption, risk factors, nodule size, or nodule location. CONCLUSIONS: Patients were generally satisfied with receiving their LCS pulmonary nodule results via letters, reporting that the letters included adequate information about their diagnosis and follow-up steps. This may provide a basis for feasible result communication via letters for cancer screening programs in underdeveloped regions in China.

In Vivo PET Detection of Lung Micrometastasis in Mice by Targeting Endothelial VCAM-1 Using a Dual-Contrast PET/MRI Probe.

Current clinical diagnostic imaging methods for lung metastases are sensitive only to large tumours (1-2 mm cross-sectional diameter), and early detection can dramatically improve treatment. We have previously demonstrated that an antibody-targeted MRI contrast agent based on microparticles of iron oxide (MPIO; 1 µm diameter) enables the imaging of endothelial vascular cell adhesion molecule-1 (VCAM-1). Using a mouse model of lung metastasis, upregulation of endothelial VCAM-1 expression was demonstrated in micrometastasis-associated vessels but not in normal lung tissue, and binding of VCAM-MPIO to these vessels was evident histologically. Owing to the lack of proton MRI signals in the lungs, we modified the VCAM-MPIO to include zirconium-89 (89Zr, t1/2 = 78.4 h) in order to allow the in vivo detection of lung metastases by positron emission tomography (PET). Using this new agent (89Zr-DFO-VCAM-MPIO), it was possible to detect the presence of micrometastases within the lung in vivo from ca. 140 µm in diameter. Histological analysis combined with autoradiography confirmed the specific binding of the agent to the VCAM-1 expressing vasculature at the sites of pulmonary micrometastases. By retaining the original VCAM-MPIO as the basis for this new molecular contrast agent, we have created a dual-modality (PET/MRI) agent for the concurrent detection of lung and brain micrometastases.

Quantitative assessment of intertarget position variations based on 4D-CT and 4D-CBCT simulations in single-isocenter multitarget lung stereotactic body radiation therapy.

BACKGROUND: In single-isocenter multitarget stereotactic body radiotherapy (SBRT), geometric miss risks arise from uncertainties in intertarget position. However, its assessment is inadequate, and may be interfered by the reconstructed tumor position errors (RPEs) during simulated CT and cone beam CT (CBCT) acquisition. This study aimed to quantify intertarget position variations and assess factors influencing it. METHODS: We analyzed data from 14 patients with 100 tumor pairs treated with single-isocenter SBRT. Intertarget position variation was measured using 4D-CT simulation to assess the intertarget position variations (ΔD) during routine treatment process. Additionally, a homologous 4D-CBCT simulation provided RPE-free comparison to determine the impact of RPEs, and isolating purely tumor motion induced ΔD to evaluate potential contributing factors. RESULTS: The median ΔD was 4.3 mm (4D-CT) and 3.4 mm (4D-CBCT). Variations exceeding 5 mm and 10 mm were observed in 31.1% and 5.5% (4D-CT) and 20.4% and 3.4% (4D-CBCT) of fractions, respectively. RPEs necessitated an additional 1-2 mm safety margin. Intertarget distance and breathing amplitude variability showed weak correlations with variation (Rs = 0.33 and 0.31). The ΔD differed significantly by locations (upper vs. lower lobe and right vs. Left lung). Notably, left lung tumor pairs exhibited the highest risk. CONCLUSIONS: This study provide a reliable way to assess intertarget position variation by using both 4D-CT and 4D-CBCT simulation. Consequently, single-isocenter SBRT for multiple lung tumors carries high risk of geometric miss. Tumor motion and RPE constitute a substantial portion of intertarget position variation, requiring correspondent strategies to minimize the intertarget uncertainties.

Differentiation of granulomatous nodules with lobulation and spiculation signs from solid lung adenocarcinomas using a CT deep learning model.

BACKGROUND: The diagnosis of solitary pulmonary nodules has always been a difficult and important point in clinical research, especially granulomatous nodules (GNs) with lobulation and spiculation signs, which are easily misdiagnosed as malignant tumors. Therefore, in this study, we utilised a CT deep learning (DL) model to distinguish GNs with lobulation and spiculation signs from solid lung adenocarcinomas (LADCs), to improve the diagnostic accuracy of preoperative diagnosis. METHODS: 420 patients with pathologically confirmed GNs and LADCs from three medical institutions were retrospectively enrolled. The regions of interest in non-enhanced CT (NECT) and venous contrast-enhanced CT (VECT) were identified and labeled, and self-supervised labels were constructed. Cases from institution 1 were randomly divided into a training set (TS) and an internal validation set (IVS), and cases from institutions 2 and 3 were treated as an external validation set (EVS). Training and validation were performed using self-supervised transfer learning, and the results were compared with the radiologists' diagnoses. RESULTS: The DL model achieved good performance in distinguishing GNs and LADCs, with area under curve (AUC) values of 0.917, 0.876, and 0.896 in the IVS and 0.889, 0.879, and 0.881 in the EVS for NECT, VECT, and non-enhanced with venous contrast-enhanced CT (NEVECT) images, respectively. The AUCs of radiologists 1, 2, 3, and 4 were, respectively, 0.739, 0.783, 0.883, and 0.901 in the (IVS) and 0.760, 0.760, 0.841, and 0.844 in the EVS. CONCLUSIONS: A CT DL model showed great value for preoperative differentiation of GNs with lobulation and spiculation signs from solid LADCs, and its predictive performance was higher than that of radiologists.

GWO+RuleFit: rule-based explainable machine-learning combined with heuristics to predict mid-treatment FDG PET response to chemoradiation for locally advanced non-small cell lung cancer.

Objective.Vital rules learned from fluorodeoxyglucose positron emission tomography (FDG-PET) radiomics of tumor subregional response can provide clinical decision support for precise treatment adaptation. We combined a rule-based machine learning (ML) model (RuleFit) with a heuristic algorithm (gray wolf optimizer, GWO) for mid-chemoradiation FDG-PET response prediction in patients with locally advanced non-small cell lung cancer.Approach.Tumors subregions were identified using K-means clustering. GWO+RuleFit consists of three main parts: (i) a random forest is constructed based on conventional features or radiomic features extracted from tumor regions or subregions in FDG-PET images, from which the initial rules are generated; (ii) GWO is used for iterative rule selection; (iii) the selected rules are fit to a linear model to make predictions about the target variable. Two target variables were considered: a binary response measure (ΔSUVmean ⩾ 20% decline) for classification and a continuous response measure (ΔSUVmean) for regression. GWO+RuleFit was benchmarked against common ML algorithms and RuleFit, with leave-one-out cross-validated performance evaluated by the area under the receiver operating characteristic curve (AUC) in classification and root-mean-square error (RMSE) in regression.Main results.GWO+RuleFit selected 15 rules from the radiomic feature dataset of 23 patients. For treatment response classification, GWO+RuleFit attained numerically better cross-validated performance than RuleFit across tumor regions and sets of features (AUC: 0.58-0.86 vs. 0.52-0.78,p= 0.170-0.925). GWO+Rulefit also had the best or second-best performance numerically compared to all other algorithms for all conditions. For treatment response regression prediction, GWO+RuleFit (RMSE: 0.162-0.192) performed better numerically for low-dimensional models (p= 0.097-0.614) and significantly better for high-dimensional models across all tumor regions except one (RMSE: 0.189-0.219,p< 0.004).Significance. The GWO+RuleFit selected rules were interpretable, highlighting distinct radiomic phenotypes that modulated treatment response. GWO+Rulefit achieved parsimonious models while maintaining utility for treatment response prediction, which can aid clinical decisions for patient risk stratification, treatment selection, and biologically driven adaptation. Clinical trial: NCT02773238.

Explainable lung cancer classification with ensemble transfer learning of VGG16, Resnet50 and InceptionV3 using grad-cam.

Medical imaging stands as a critical component in diagnosing various diseases, where traditional methods often rely on manual interpretation and conventional machine learning techniques. These approaches, while effective, come with inherent limitations such as subjectivity in interpretation and constraints in handling complex image features. This research paper proposes an integrated deep learning approach utilizing pre-trained models-VGG16, ResNet50, and InceptionV3-combined within a unified framework to improve diagnostic accuracy in medical imaging. The method focuses on lung cancer detection using images resized and converted to a uniform format to optimize performance and ensure consistency across datasets. Our proposed model leverages the strengths of each pre-trained network, achieving a high degree of feature extraction and robustness by freezing the early convolutional layers and fine-tuning the deeper layers. Additionally, techniques like SMOTE and Gaussian Blur are applied to address class imbalance, enhancing model training on underrepresented classes. The model's performance was validated on the IQ-OTH/NCCD lung cancer dataset, which was collected from the Iraq-Oncology Teaching Hospital/National Center for Cancer Diseases over a period of three months in fall 2019. The proposed model achieved an accuracy of 98.18%, with precision and recall rates notably high across all classes. This improvement highlights the potential of integrated deep learning systems in medical diagnostics, providing a more accurate, reliable, and efficient means of disease detection.

The utility of artificial intelligence in identifying radiological evidence of lung cancer and pulmonary tuberculosis in a high-burden tuberculosis setting.

BACKGROUND: Artificial intelligence (AI), using deep learning (DL) systems, can be utilised to detect radiological changes of various pulmonary diseases. Settings with a high burden of tuberculosis (TB) and people living with HIV can potentially benefit from the use of AI to augment resource-constrained healthcare systems. OBJECTIVE: To assess the utility of qXR software (AI) in detecting radiological changes compatible with lung cancer or pulmonary TB (PTB). METHODS: We performed an observational study in a tertiary institution that serves a population with a high burden of lung cancer and PTB. In total, 382 chest radiographs that had a confirmed diagnosis were assessed: 127 with lung cancer, 144 with PTB and 111 normal. These chest radiographs were de-identified and randomly uploaded by a blinded investigator into qXR software. The output was generated as probability scores from predefined threshold values. RESULTS: The overall sensitivity of the qXR in detecting lung cancer was 84% (95% confidence interval (CI) 80 - 87%), specificity 91% (95% CI 84 - 96%) and positive predictive value of 97% (95% CI 95 - 99%). For PTB, it had a sensitivity of 90% (95% CI 87 - 93%) and specificity of 79% (95% CI 73 - 84%) and negative predictive value of 85% (95% CI 79 - 91%). CONCLUSION: The qXR software was sensitive and specific in categorising chest radiographs as consistent with lung cancer or TB, and can potentially aid in the earlier detection and management of these diseases.

Combined expert-in-the-loop-random forest multiclass segmentation U-net based artificial intelligence model: evaluation of non-small cell lung cancer in fibrotic and non-fibrotic microenvironments.

BACKGROUND: The tumor microenvironment (TME) plays a key role in lung cancer initiation, proliferation, invasion, and metastasis. Artificial intelligence (AI) methods could potentially accelerate TME analysis. The aims of this study were to (1) assess the feasibility of using hematoxylin and eosin (H&E)-stained whole slide images (WSI) to develop an AI model for evaluating the TME and (2) to characterize the TME of adenocarcinoma (ADCA) and squamous cell carcinoma (SCCA) in fibrotic and non-fibrotic lung. METHODS: The cohort was derived from chest CT scans of patients presenting with lung neoplasms, with and without background fibrosis. WSI images were generated from slides of all 76 available pathology cases with ADCA (n = 53) or SCCA (n = 23) in fibrotic (n = 47) or non-fibrotic (n = 29) lung. Detailed ground-truth annotations, including of stroma (i.e., fibrosis, vessels, inflammation), necrosis and background, were performed on WSI and optimized via an expert-in-the-loop (EITL) iterative procedure using a lightweight [random forest (RF)] classifier. A convolution neural network (CNN)-based model was used to achieve tissue-level multiclass segmentation. The model was trained on 25 annotated WSI from 13 cases of ADCA and SCCA within and without fibrosis and then applied to the 76-case cohort. The TME analysis included tumor stroma ratio (TSR), tumor fibrosis ratio (TFR), tumor inflammation ratio (TIR), tumor vessel ratio (TVR), tumor necrosis ratio (TNR), and tumor background ratio (TBR). RESULTS: The model's overall classification for precision, sensitivity, and F1-score were 94%, 90%, and 91%, respectively. Statistically significant differences were noted in TSR (p = 0.041) and TFR (p = 0.001) between fibrotic and non-fibrotic ADCA. Within fibrotic lung, statistically significant differences were present in TFR (p = 0.039), TIR (p = 0.003), TVR (p = 0.041), TNR (p = 0.0003), and TBR (p = 0.020) between ADCA and SCCA. CONCLUSION: The combined EITL-RF CNN model using only H&E WSI can facilitate multiclass evaluation and quantification of the TME. There are significant differences in the TME of ADCA and SCCA present within or without background fibrosis. Future studies are needed to determine the significance of TME on prognosis and treatment.

Habitat radiomics and deep learning fusion nomogram to predict EGFR mutation status in stage I non-small cell lung cancer: a multicenter study.

Develop a radiomics nomogram that integrates deep learning, radiomics, and clinical variables to predict epidermal growth factor receptor (EGFR) mutation status in patients with stage I non-small cell lung cancer (NSCLC). We retrospectively included 438 patients who underwent curative surgery and completed driver-gene mutation tests for stage I NSCLC from four academic medical centers. Predictive models were established by extracting and analyzing radiomic features in intratumoral, peritumoral, and habitat regions of CT images to identify EGFR mutation status in stage I NSCLC. Additionally, three deep learning models based on the intratumoral region were constructed. A nomogram was developed by integrating representative radiomic signatures, deep learning, and clinical features. Model performance was assessed by calculating the area under the receiver operating characteristic (ROC) curve. The established habitat radiomics features demonstrated encouraging performance in discriminating between EGFR mutant and wild-type, with predictive ability superior to other single models (AUC 0.886, 0.812, and 0.790 for the training, validation, and external test sets, respectively). The radiomics-based nomogram exhibited excellent performance, achieving the highest AUC values of 0.917, 0.837, and 0.809 in the training, validation, and external test sets, respectively. Decision curve analysis (DCA) indicated that the nomogram provided a higher net benefit than other radiomics models, offering valuable information for treatment.