Machine learning-based histological classification that predicts recurrence of peripheral lung squamous cell carcinoma

A novel histopathological feature of spatial tumor-stroma distribution predicts lung squamous cell carcinoma prognosis

We used a mathematical approach to investigate the quantitative spatial profile of cancer cells and stroma in lung squamous cell carcinoma tissues and its clinical relevance. The study enrolled 132 patients with 3-5 cm peripheral lung squamous cell carcinoma, resected at the National Cancer Center Hospital East. We utilized machine learning to segment cancer cells and stroma on cytokeratin AE1/3 immunohistochemistry images. Subsequently, a spatial form of Shannon's entropy was employed to precisely quantify the spatial distribution of cancer cells and stroma. This quantification index was defined as the spatial tumor-stroma distribution index (STSDI). The patients were classified as STSDI-low and -high groups for clinicopathological comparison. The STSDI showed no significant association with baseline clinicopathological features, including sex, age, pathological stage, and lymphovascular invasion. However, the STSDI-low group had significantly shorter recurrence-free survival (5-years RFS: 49.5% vs. 76.2%, p < 0.001) and disease-specific survival (5-years DSS: 53.6% vs. 81.5%, p < 0.001) than the STSDI-high group. In contrast, the application of Shannon's entropy without spatial consideration showed no correlation with patient outcomes. Moreover, low STSDI was an independent unfavorable predictor of tumor recurrence and disease-specific death (RFS; HR = 2.668, p < 0.005; DSS; HR = 3.057, p < 0.005), alongside the pathological stage. Further analysis showed a correlation between low STSDI and destructive growth patterns of cancer cells within tumors, potentially explaining the aggressive nature of STSDI-low tumors. In this study, we presented a novel approach for histological analysis of cancer tissues that revealed the prognostic significance of spatial tumor-stroma distribution in lung squamous cell carcinoma.

Establishment of artificial intelligence model for precise histological subtyping of lung adenocarcinoma and its application to quantitative and spatial analysis

BACKGROUND

The histological subtype of lung adenocarcinoma is a major prognostic factor. We developed a new artificial intelligence model to classify lung adenocarcinoma images into seven histological subtypes and adopted the model for whole-slide images to investigate the relationship between the distribution of histological subtypes and clinicopathological factors.

METHODS

Using histological subtype images, which are typical for pathologists, we trained and validated an artificial intelligence model. Then, the model was applied to whole-slide images of resected lung adenocarcinoma specimens from 147 cases.

RESULT

The model achieved an accuracy of 99.7% in training sets and 90.4% in validation sets consisting of typical tiles of histological subtyping for pathologists. When the model was applied to whole-slide images, the predominant subtype according to the artificial intelligence model classification matched that determined by pathologists in 75.5% of cases. The predominant subtype and tumor grade (using the WHO fourth and fifth classifications) determined by the artificial intelligence model resulted in similar recurrence-free survival curves to those determined by pathologists. Furthermore, we stratified the recurrence-free survival curves for patients with different proportions of high-grade components (solid, micropapillary and cribriform) according to the physical distribution of the high-grade component. The results suggested that tumors with centrally located high-grade components had a higher malignant potential (P < 0.001 for 5-20% high-grade component).

CONCLUSION

The new artificial intelligence model for histological subtyping of lung adenocarcinoma achieved high accuracy, and subtype quantification and subtype distribution analyses could be achieved. Artificial intelligence model therefore has potential for clinical application for both quantification and spatial analysis.

The proportion of tumour stroma predicts response to treatment of immune checkpoint inhibitor in combination with chemotherapy in patients with stage IIIB-IV non-small cell lung cancer

AIMS

Immunotherapy has brought a new era to cancer treatment, yet we lack dependable predictors for its effectiveness. This study explores the predictive significance of intratumour stroma proportion (iTSP) for treatment success and prognosis in non-small cell lung cancer (NSCLC) patients undergoing treatment with immune check-point inhibitors (ICIs) together with chemotherapy.

METHODS AND RESULTS

We retrospectively collected data from patients with unresectable stage IIIB-IV NSCLC who were treated with first-line ICIs and chemotherapy. Each patient received a confirmed pathological diagnosis, and the pathologist evaluated the iTSP on haematoxylin and eosin (H&E)-stained sections of diagnostic tissue slides. Among the 102 H&E-stained biopsy samples, 61 (59.8%) were categorised as stroma-L (less than 50% iTSP), while 41 (40.2%) were classified as stroma-H (more than 50% iTSP). We observed that the stroma-L group exhibited a significantly better objective response rate (ORR) (72.1 versus 51.2%, P = 0.031) and deeper response depth (DpR) (-50.49 ± 28.79% versus -35.83 ± 29.91%, P = 0.015) compared to the stroma-H group. Furthermore, the stroma-L group showed longer median progression-free survival (PFS) (9.6 versus 6.0 months, P = 0.011) and overall survival (OS) (24.0 versus 12.2 months, P = 0.001) compared to the stroma-H group. Multivariate Cox proportional hazards regression analysis indicated that iTSP was a highly significant prognostic factor for both PFS [hazard ratio (HR) = 1.713; P = 0.030] and OS (HR = 2.225; P = 0.003).

CONCLUSION

Our findings indicate that a lower iTSP corresponds to improved clinical outcomes and greater DpR in individuals with stage IIIB-IV NSCLC treated with first-line ICIs and chemotherapy. The iTSP could potentially serve as a predictive biomarker for ICIs therapy response.

Insights into metastatic roadmap of head and neck cancer squamous cell carcinoma based on clinical, histopathological and molecular profiles

The incidence of head and neck cancer (HNC), constituting approximately one in ten cancer cases worldwide, affects approximately 644,000 individuals annually. Managing this complex disease involves various treatment modalities such as systemic therapy, radiation, and surgery, particularly for patients with locally advanced disease. HNC treatment necessitates a multidisciplinary approach due to alterations in patients' genomes affecting their functionality. Predominantly, squamous cell carcinomas (SCCs), the majority of HNCs, arise from the upper aerodigestive tract epithelium. The epidemiology, staging, diagnosis, and management techniques of head and neck squamous cell carcinoma (HNSCC), encompassing clinical, image-based, histopathological and molecular profiling, have been extensively reviewed. Lymph node metastasis (LNM) is a well-known predictive factor for HNSCC that initiates metastasis and significantly impacts HNSCC prognosis. Distant metastasis (DM) in HNSCC has been correlated to aberrant expression of cancer cell-derived cytokines and growth factors triggering abnormal activation of several signaling pathways that boost cancer cell aggressiveness. Recent advances in genetic profiling, understanding tumor microenvironment, oligometastatic disease, and immunotherapy have revolutionized treatment strategies and disease control. Future research may leverage genomics and proteomics to identify biomarkers aiding individualized HNSCC treatment. Understanding the molecular basis, genetic landscape, atypical signaling pathways, and tumor microenvironment have enhanced the comprehension of HNSCC molecular etiology. This critical review sheds light on regional and distant metastases in HNSCC, presenting major clinical and laboratory features, predictive biomarkers, and available therapeutic approaches.

Construction of a predictive model for bone metastasis from first primary lung adenocarcinoma within 3 cm based on machine learning algorithm: a retrospective study

Background

Adenocarcinoma, the most prevalent histological subtype of non-small cell lung cancer, is associated with a significantly higher likelihood of bone metastasis compared to other subtypes. The presence of bone metastasis has a profound adverse impact on patient prognosis. However, to date, there is a lack of accurate bone metastasis prediction models. As a result, this study aims to employ machine learning algorithms for predicting the risk of bone metastasis in patients.

Method

We collected a dataset comprising 19,454 cases of solitary, primary lung adenocarcinoma with pulmonary nodules measuring less than 3 cm. These cases were diagnosed between 2010 and 2015 and were sourced from the Surveillance, Epidemiology, and End Results (SEER) database. Utilizing clinical feature indicators, we developed predictive models using seven machine learning algorithms, namely extreme gradient boosting (XGBoost), logistic regression (LR), light gradient boosting machine (LightGBM), Adaptive Boosting (AdaBoost), Gaussian Naive Bayes (GNB), multilayer perceptron (MLP) and support vector machine (SVM).

Results

The results demonstrated that XGBoost exhibited superior performance among the four algorithms (training set: AUC: 0.913; test set: AUC: 0.853). Furthermore, for convenient application, we created an online scoring system accessible at the following URL: https://www.xsmartanalysis.com/model/predict/?mid=731symbol=7Fr16wX56AR9Mk233917, which is based on the highest performing model.

Conclusion

XGBoost proves to be an effective algorithm for predicting the occurrence of bone metastasis in patients with solitary, primary lung adenocarcinoma featuring pulmonary nodules below 3 cm in size. Moreover, its robust clinical applicability enhances its potential utility.

Detection of Changes in CEA and ProGRP Levels in BALF of Patients with Peripheral Lung Cancer and the Relationship with CT Signs

Objective

To investigate the relationship between the detection of changes in the levels of carcinoembryonic antigen (CEA) and progastrin-releasing peptide (ProGRP) in bronchoalveolar lavage fluid (BALF) and CT signs in patients with peripheral lung cancer.

Methods

Retrospective analysis of 108 patients with perihilar lung cancer who attended our hospital from January 2019 to January 2022, 54 cases were randomly selected as the observation group and 50 cases as the control group. Patients in both groups received CT examination and BALF test at the same time to observe and compare the differences in serum levels, the relationship between CT signs and serum indices, and the diagnostic value of peripheral lung cancer between the two groups.

Results

The serum levels of ProGrp, CEA, CA211, and NSE in the observation group were significantly higher than those in the control group, and the difference was statistically significant (P < 0.05). The morphology, density, mass enhancement pattern, bronchial morphology, obstructive signs, and lymph node fusion of CT signs were compared between the observation group and the control group, indicating that CT signs were more helpful for the localization, diagnosis, and staging of lung cancer. The results of ROC curve analysis showed that the AUC value of low-dose CT combined with serum ProGrp, CEA, CA211, and NSE was 0.892, sensitivity was 96.21%, and specificity of 90.05%, which were significantly higher than those of the single tests, respectively. The positive likelihood ratio was 84.41% and the negative likelihood ratio was 87.11%.

Conclusion

The combination of CT signs and serum tumour markers helps to improve the detection rate, sensitivity, and specificity of lung cancer, which has a high diagnostic rate for lung cancer and may provide evidence for the early diagnosis of lung cancer.

Novel cellular senescence-related risk model identified as the prognostic biomarkers for lung squamous cell carcinoma

Background

Lung cancer is one of the top causes of cancer-related death worldwide. Cellular senescence is a characteristic of cell cycle arrest that plays a role in carcinogenesis and immune microenvironment modulation. Despite this, the clinical and immune cell infiltration features of senescence in lung squamous cell carcinoma (LUSC) are unknown.

Methods

The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) were used to get RNA-seq data and clinical information for LUSC. The least absolute shrinkage and selection operator (LASSO)-Cox regression, receiver operating characteristic (ROC), and Kaplan-Meier analysis were used to evaluate a risk model for predicting overall survival based on six differentially expressed genes. The tumor microenvironment (TME) and immunotherapy response were also studied.

Results

To discriminate LUSC into high- and low-risk subgroups, a risk model comprised of six cellular senescence-related genes (CDKN1A, CEBPB, MDH1, SIX1, SNAI1, and SOX5) was developed. The model could stratify patients into high-risk and low-risk groups, according to ROC and Kaplan-Meier analysis. In the TCGA-LUSC and GSE73403 cohorts, the high-risk group had a worse prognosis (P&lt;0.05), and was associated with immune cell inactivation and being insensitive to immunotherapy in IMvigor210.

Conclusions

We discovered a new LUSC classification based on six cellular senescence-related genes, which will aid in identifying patients who will benefit from anti-PD-1 treatment. Targeting senescence-related genes appears to be another option for improving clinical therapy for LUSC.

Prognostic Value of Selected Histologic Features for Lung Squamous Cell Carcinoma

The recent histologic subtyping of lung adenocarcinoma has demonstrated the prognostic values of histologic patterns in this malignancy. However, the histological features of lung squamous cell carcinoma (SCC) are much less established. This short review discusses several promising histological prognostic markers for SCC, including tumor budding, tumor cell nesting, and the spreading of tumors through air spaces. Wherever appropriate, the biological significance of these morphological features was also discussed. The investigators consider that histological prognostic markers are highly valuable in understanding the cancer biology of SCC, and in guiding clinical treatment. However, larger clinical cohorts are needed to better establish the prognostic values of the aforementioned histological markers. The application of modern technologies, including machine-learning, would make the histological analysis accurate and reproducible.

Fusing pre-trained convolutional neural networks features for multi-differentiated subtypes of liver cancer on histopathological images

Liver cancer is a malignant tumor with high morbidity and mortality, which has a tremendous negative impact on human survival. However, it is a challenging task to recognize tens of thousands of histopathological images of liver cancer by naked eye, which poses numerous challenges to inexperienced clinicians. In addition, factors such as long time-consuming, tedious work and huge number of images impose a great burden on clinical diagnosis. Therefore, our study combines convolutional neural networks with histopathology images and adopts a feature fusion approach to help clinicians efficiently discriminate the differentiation types of primary hepatocellular carcinoma histopathology images, thus improving their diagnostic efficiency and relieving their work pressure. In this study, for the first time, 73 patients with different differentiation types of primary liver cancer tumors were classified. We performed an adequate classification evaluation of liver cancer differentiation types using four pre-trained deep convolutional neural networks and nine different machine learning (ML) classifiers on a dataset of liver cancer histopathology images with multiple differentiation types. And the test set accuracy, validation set accuracy, running time with different strategies, precision, recall and F1 value were used for adequate comparative evaluation. Proved by experimental results, fusion networks (FuNet) structure is a good choice, which covers both channel attention and spatial attention, and suppresses channel interference with less information. Meanwhile, it can clarify the importance of each spatial location by learning the weights of different locations in space, then apply it to the study of classification of multi-differentiated types of liver cancer. In addition, in most cases, the Stacking-based integrated learning classifier outperforms other ML classifiers in the classification task of multi-differentiation types of liver cancer with the FuNet fusion strategy after dimensionality reduction of the fused features by principle component analysis (PCA) features, and a satisfactory result of 72.46% is achieved in the test set, which has certain practicality.

Predictive Role of Tumor-Stroma Ratio for Survival of Patients With Non-Small Cell Lung Cancer: A Meta-Analysis

Background: Role of tumor-stroma ratio (TSR) as a predictor of survival in patients with non-small cell lung cancer (NSCLC) remains not clear. A systematic review and meta-analysis was conducted to summarize current evidence for the role of TSR in NSCLC.

Methods: Relevant cohort studies were retrieved via search of Medline, Embase, and Web of Science databases. The data was combined with a random-effect model by incorporating the between-study heterogeneity. Specifically, subgroup and meta-regression analyses were performed to explore the association between TSR and survival in patients with squamous cell carcinoma (SCC) or adenocarcinoma (AC).

Results: Nine cohort studies with 2031 patients with NSCLC were eligible for the meta-analysis. Pooled results showed that compared to those stroma-poor tumor, patients with stroma rich NSCLC were associated with worse recurrence-free survival (RFS, hazard ratio [HR] = 1.52, 95% confidence interval [CI]: 1.07 to 2.16, p = 0.02) and overall survival (OS, HR = 1.48, 95% CI: 1.20 to 1.82, p < 0.001). Subgroup analyses showed that stroma-rich tumor may be associated with a worse survival of SCC (HR = 1.89 and 1.47 for PFS and OS), but a possibly favorable survival of AC (HR = 0.28 and 0.69 for PFS and OS). Results of meta-regression analysis also showed that higher proportion of patients with SCC was correlated with higher HRs for RFS (Coefficient = 0.012, p = 0.03) and OS (Coefficient = 0.014, p = 0.02) in the included patients, while higher proportion of patients with AC was correlated with lower HRs for RFS (Coefficient = −0.012, p = 0.03) and OS (Coefficient = −0.013, p = 0.04), respectively.

Conclusion: Tumor TSR could be used as a predictor of survival in patients with NSCLC. The relative proportion of patients with SCC/AC in the included NSCLC patients may be an important determinant for the association between TSR and survival in NSCLC. Stroma richness may be a predictor of poor survival in patients with lung SCC, but a predictor of better survival in patients with lung AC.

Artificial intelligence in thoracic surgery: a narrative review

OBJECTIVE

The aim of this article is to review the current applications of artificial intelligence in thoracic surgery, from diagnosis and pulmonary disease management, to preoperative risk-assessment, surgical planning, and outcomes prediction.

BACKGROUND

Artificial intelligence implementation in healthcare settings is rapidly growing, though its widespread use in clinical practice is still limited. The employment of machine learning algorithms in thoracic surgery is wide-ranging, including all steps of the clinical pathway.

METHODS

We performed a narrative review of the literature on Scopus, PubMed and Cochrane databases, including all the relevant studies published in the last ten years, until March 2021.

CONCLUSION

Machine learning methods are promising encouraging results throughout the key issues of thoracic surgery, both clinical, organizational, and educational. Artificial intelligence-based technologies showed remarkable efficacy to improve the perioperative evaluation of the patient, to assist the decision-making process, to enhance the surgical performance, and to optimize the operating room scheduling. Still, some concern remains about data supply, protection, and transparency, thus further studies and specific consensus guidelines are needed to validate these technologies for daily common practice.

KEYWORDS

Artificial intelligence (AI); thoracic surgery; machine learning; lung resection; perioperative medicine.

Deep learning based microscopic cell images classification framework using multi-level ensemble

BACKGROUND AND OBJECTIVES

Advancement of the ultra-fast microscopic images acquisition and generation techniques give rise to the automated artificial intelligence (AI)-based microscopic images classification systems. The earlier cell classification systems classify the cell images of a specific type captured using a specific microscopy technique, therefore the motivation behind the present study is to develop a generic framework that can be used for the classification of cell images of multiple types captured using a variety of microscopic techniques.

METHODS

The proposed framework for microscopic cell images classification is based on the transfer learning-based multi-level ensemble approach. The ensemble is made by training the same base model with different optimisation methods and different learning rates. An important contribution of the proposed framework lies in its ability to capture different granularities of features extracted from multiple scales of an input microscopic cell image. The base learners used in the proposed ensemble encapsulates the aggregation of low-level coarse features and high-level semantic features, thus, represent the different granular microscopic cell image features present at different scales of input cell images. The batch normalisation layer has been added to the base models for the fast convergence in the proposed ensemble for microscopic cell images classification.

RESULTS

The general applicability of the proposed framework for microscopic cell image classification has been tested with five different public datasets. The proposed method has outperformed the experimental results obtained in several other similar works.

CONCLUSIONS

The proposed framework for microscopic cell classification outperforms the other state-of-the-art classification methods in the same domain with a comparatively lesser amount of training data.

Intrapulmonic Cavity or Necrosis on Baseline CT Scan Serves as an Efficacy Predictor of Anti-PD-(L)1 Inhibitor in Advanced Lung Squamous Cell Carcinoma

Background

Predictive markers for guidance and monitoring of immunotherapy in lung squamous cell carcinoma (LSCC) are an interesting topic but have yet to be fully explored. A primary characteristic of LSCC is tumor necrosis that results in extensive immune suppression in patients. We sought to assess whether tumor necrosis or cavity on baseline CT could effectively predict the efficacy of immune checkpoint inhibitors (ICIs) in advanced LSCC.

Methods

Advanced LSCC cases undergoing pre-treatment chest CT imaging and receiving ICIs were retrospectively collected. All CT images were reviewed by an independent chest radiologist blinded to any previous diagnosis to confirm morphological alterations in necrosis or cavity. We performed Logistic regression and developed Cox proportional hazards models to assess the predictive performance of baseline necrosis or cavity characteristics in advanced LSCC. Survival estimates were observed using Kaplan–Meier curves.

Results

Ninety-three patients were eligible for analysis, predominantly consisting of patients with ECOG performance status of 0 or 1 (97.8%), male patients (95.7%), and heavy smokers (92.5%). Intrapulmonic necrosis or cavity on CT scan was present in 52.7% of all patients. Generally, the objective response rate (ORR) in patients with necrosis or cavity to ICI treatment was significantly worse versus those without (30.6% vs 54.5%, p = 0.020), with the subgroup ORRs as follows: ICI monotherapy (necrosis vs non-necrosis: 10.0% vs 36.8%, p =0.047) and ICI combination therapy (44.8% vs 68.0%, p =0.088). Multivariable analysis identified intrapulmonic necrosis or cavity at baseline as a major risk factor for advanced LSCC (HR 4.042, 95% CI1.149–10.908, p = 0.006). Multivariate Cox analysis showed that baseline necrosis or cavity and ICI monotherapy were unfavorable factors for progression-free survival (HR 1.729; 95% CI1.203–2.484, p =0.003).

Conclusion

LSCC patients with intrapulmonic cavity or necrosis on baseline CT scan may respond poorly to anti-PD-(L)1-treatment, monotherapy and combination therapy alike.

The prognostic value of the tumor-stroma ratio in squamous cell lung cancer, a cohort study

OBJECTIVES

The tumor-stroma ratio (TSR) is based on the relative amount of stroma in the primary tumor and has proven to be an independent prognostic factor in various solid tumors. The prognosis of patients and adjuvant treatment decision making in lung squamous cell carcinomas (SqCC) is based on the TNM classification. Currently, no other prognostic biomarkers are available. In this study we evaluated the prognostic value of the TSR in lung SqCC.

MATERIAL AND METHODS

Patients undergoing lung surgery because of lung SqCC between 2000 and 2018 at the Leiden University Medical Center were included. The TSR was scored on hematoxylin & eosin stained tissue sections. Based on the amount of tumor-stroma, two groups were defined: ≤50% was classified as a stroma-low tumor and >50% as stroma-high. The prognostic value of the TSR was determined with survival analysis.

RESULTS

A total of 174 stage I-III patients were included. Of them, 79 (45%) were stroma-low and 95 (55%) stroma-high. Separately analyzed for tumor stages, the TSR showed to be an independent prognostic biomarker in stage II (n = 68) for 5-year overall survival (HR=3.0; 95% CI, 1.1-8.6; p = 0.035) and 5-year disease free survival (DFS) (HR=3.6; 95% CI, 1.3-9.9; p = 0.014). Patients with a stroma-high tumor had a worse 5-year DFS in the whole cohort (HR 1.6; 95% CI, 1.0-2.4; p = 0.048), but no independent prognostic value was found.

CONCLUSION

In stage II lung SqCC patients, stroma-low tumors have a better prognosis compared to stroma-high tumors. Moreover, adjuvant chemotherapy could be spared for these stroma-low patients.