CT imaging-based machine learning model: a potential modality for predicting low-risk and high-risk groups of thymoma: "Impact of surgical modality choice".

INTRODUCTION: Radiomics methods are used to analyze various medical images, including computed tomography (CT), magnetic resonance, and positron emission tomography to provide information regarding the diagnosis, patient outcome, tumor phenotype, and the gene-protein signatures of various diseases. In low-risk group, complete surgical resection is typically sufficient, whereas in high-risk thymoma, adjuvant therapy is usually required. Therefore, it is important to distinguish between both. This study evaluated the CT radiomics features of thymomas to discriminate between low- and high-risk thymoma groups. MATERIALS AND METHODS: In total, 83 patients with thymoma were included in this study between 2004 and 2019. We used the Radcloud platform (Huiying Medical Technology Co., Ltd.) to manage the imaging and clinical data and perform the radiomics statistical analysis. The training and validation datasets were separated by a random method with a ratio of 2:8 and 502 random seeds. The histopathological diagnosis was noted from the pathology report. RESULTS: Four machine-learning radiomics features were identified to differentiate a low-risk thymoma group from a high-risk thymoma group. The radiomics feature names were Energy, Zone Entropy, Long Run Low Gray Level Emphasis, and Large Dependence Low Gray Level Emphasis. CONCLUSIONS: The results demonstrated that a machine-learning model and a multilayer perceptron classifier analysis can be used on CT images to predict low- and high-risk thymomas. This combination could be a useful preoperative method to determine the surgical approach for thymoma.

Microcomputed tomography as a diagnostic tool for detection of lymph node metastasis in non-small cell lung cancer: A decision-support approach for pathological examination "A pilot study for method validation".

Background: Non-small cell lung cancer (NSCLC) patients without lymph node (LN) metastases (pN0) may exhibit different survival rates, even when their T stage is similar. This divergence could be attributed to the current pathology practice, wherein LNs are examined solely in two-dimensional (2D). Unfortunately, adhering to the protocols of 2D pathological examination does not ensure the exhaustive sampling of all excised LNs, thereby leaving room for undetected metastatic foci in the unexplored depths of tissues. The employment of micro-computed tomography (micro-CT) facilitates a three-dimensional (3D) evaluation of all LNs without compromising sample integrity. In our study, we utilized quantitative micro-CT parameters to appraise the metastatic status of formalin-fixed paraffin-embedded (FFPE) LNs. Methods: Micro-CT scans were conducted on 12 FFPEs obtained from 8 NSCLC patients with histologically confirmed mediastinal LN metastases. Simultaneously, whole-slide images from these FFPEs underwent scanning, and 47 regions of interest (ROIs) (17 metastatic foci, 11 normal lymphoid tissues, 10 adipose tissues, and 9 anthracofibrosis) were marked on scanned images. Quantitative structural variables obtained via micro-CT analysis from tumoral and non-tumoral ROIs, were analyzed. Result: Significant distinctions were observed in linear density, connectivity, connectivity density, and closed porosity between tumoral and non-tumoral ROIs, as indicated by kappa coefficients of 1, 0.90, 1, and 1, respectively. Receiver operating characteristic analysis substantiated the differentiation between tumoral and non-tumoral ROIs based on thickness, linear density, connectivity, connectivity density, and the percentage of closed porosity. Conclusions: Quantitative micro-CT parameters demonstrate the ability to distinguish between tumoral and non-tumoral regions of LNs in FFPEs. The discriminatory characteristics of these quantitative micro-CT parameters imply their potential usefulness in developing an artificial intelligence algorithm specifically designed for the 3D identification of LN metastases while preserving the FFPE tissue.

Oncological results in primary and secondary malignant chest wall tumors.

Background: This study aims to evaluate the oncological results of primary and secondary chest wall tumors treated with curative resections and to investigate possible prognostic factors. Methods: Between January 2010 and December 2021, a total of 77 patients (53 males, 24 females; median age: 59 years; range, 3 to 87 years) who underwent curative resection for malignant chest wall tumors were retrospectively analyzed. Each tumor was staged according to its histological type. Age, sex, tumor diameter, tumor type (primary/secondary), histological tumor type, grade, stage, complete resection, rib resection, reconstruction, neoadjuvant and adjuvant therapy, recurrence, and survival data were recorded. Results: Of the chest wall tumors, 33 (42.9%) were primary and 44 (57.1%) were secondary (local invasion, metastasis). Nine (11.7%) patients had positive surgical margins. Chest wall resection was most commonly performed due to lung cancer invasion (46.8%), followed by Ewing sarcoma (13%). Recurrence was observed in 34 (44.2%) patients. The five-year recurrence-free survival rate was 42.7% and the five-year overall survival rate was 58.6%. There was no significant difference between the primary and secondary tumors in terms of recurrence-free and overall survival (p=0.663 and p=0.313, respectively). In the multivariate analysis, tumor grade and rib resection were found to be independent prognostic factors for both recurrence-free survival (p=0.005 and p<0.001, respectively) and overall survival (p=0.048 and p=0.007, respectively). Conclusion: Successful oncological results can be achieved in wellselected patients with primary and secondary chest wall tumors. The grade of the tumor should be taken into account while determining the neoadjuvant or adjuvant treatment approach and surgical margin width. Rib resection should not be avoided when necessary.

HER2 Amplification by Next-Generation Sequencing in Lung Carcinoma: A Comparison of NGS Amplified and Non-amplified Cases by Immunohistochemistry and In Situ Hybridization

Background: Although the role of HER2 amplification and its evaluation methods are well known in breast carcinoma, methods for detection of HER2 amplification in non-small cell lung carcinoma are unclear. Next-generation sequencing is widely used in searching multiple therapeutic targets, and it is possible to evaluate copy number variation of genes by next-generation sequencing. Aims: To re-evaluate the HER2 status of non-small cell lung carcinoma cases detected as HER2 amplified and non-amplified by next-generation sequencing via the most commonly used HER2 investigation methods in routine pathology practice, namely immunohistochemistry and in situ hybridization. Study Design: Retrospective cross-sectional study. Methods: Among the 256 patients whose mutation profiles were examined by next-generation sequencing, HER2 amplified (13 cases) and non-HER2-amplified (13 cases) were determined as study and control groups, respectively, by next-generation sequencing. HER2 next-generation sequencing amplified tumors were investigated for HER2 expression and amplification using immunohistochemistry and silver in situ hybridization. Results: From a group of 256 non-small cell lung carcinoma, 33 tumors (12.8%) showed HER2 amplification with next-generation sequencing. Although we observed more frequent HER2 positivity by immunohistochemistry in next-generation sequencing-amplified cases, when compared to non-amplified cases (50% and 23% respectively), the difference was not significant (P = .221). Within the HER2 amplified group, inter-method-agreement was very good between next-generation sequencing results amplification and in situ hybridization status. Next-generation sequencing results showed a strong interclass correlation coefficient with HER2/cell (P = .009, r = 0.777) and HER2/CEP17 ratio (P = .001, r = 0.805). The median HER2/CEP17 ratio was higher in the next-generation sequencing amplified group (P = .013); however, three cases were found to be amplified by silver in situ hybridization among the next-generation sequencing non-amplified cases. EGFR and FGFR1 amplification were more frequent in HER2 next-generation sequencing amplified group than next-generation sequencing non-amplified group (P < .001). Conclusion: Until the effects of HER2 amplification on the HER2 protein are well understood and pulmonary carcinoma algorithms are defined, non-small cell lung carcinomas found to be amplified by next-generation sequencing should be verified by additional methods.

gsem: A Stata command for parametric joint modelling of longitudinal and accelerated failure time models.

BACKGROUND: The number of studies using joint modelling of longitudinal and survival data have increased in the past two decades, but analytical techniques and software shortcomings have remained. A joint model is often used for analysis of a combination of longitudinal sub-model and survival sub-model using shared random effects. Cox regression commonly referring to the survival sub-model, should not be used when proportional hazards assumptions are not satisfied. In such cases, the parametric survival model is preferable. METHODS: We describe different parametric survival models for survival sub-model of joint modelling. We demonstrate how these models can be fit using gsem command (used for generalized structural equation model) in Stata that allows the model to be jointly continuous longitudinal and parametric survival data. With this code, linear mixed effect model is used for the longitudinal sub-model of the joint model, allowing random and fixed effects of the time. In gsem command for survival sub-models, there are five different choices: exponential, Weibull, log-normal, log-logistic and gamma accelerated failure time models. RESULTS: In this paper, we have described properties of gsem command for parametric joint modelling and have shown an application for parametric joint models on the 312 patients with primary biliary cirrhosis, which is a major health problem in the western world. CONCLUSIONS: We showed how parametric joint models can be used with the gsem command which has been the only Stata code in the literature to fit the parametric joint models, for the generalized structural equation model, and we used the primary biliary cirrhosis dataset for the detailed application of the command. Thus, the gsem command becomes more useful for fitting parametric joint models.