Multivariable Risk Modelling and Survival Analysis with Machine Learning in SARS-CoV-2 Infection.

AIM: To evaluate the performance of a machine learning model based on demographic variables, blood tests, pre-existing comorbidities, and computed tomography(CT)-based radiomic features to predict critical outcome in patients with acute respiratory syndrome coronavirus 2 (SARS-CoV-2). METHODS: We retrospectively enrolled 694 SARS-CoV-2-positive patients. Clinical and demographic data were extracted from clinical records. Radiomic data were extracted from CT. Patients were randomized to the training (80%, n = 556) or test (20%, n = 138) dataset. The training set was used to define the association between severity of disease and comorbidities, laboratory tests, demographic, and CT-based radiomic variables, and to implement a risk-prediction model. The model was evaluated using the C statistic and Brier scores. The test set was used to assess model prediction performance. RESULTS: Patients who died (n = 157) were predominantly male (66%) over the age of 50 with median (range) C-reactive protein (CRP) = 5 [1, 37] mg/dL, lactate dehydrogenase (LDH) = 494 [141, 3631] U/I, and D-dimer = 6.006 [168, 152.015] ng/mL. Surviving patients (n = 537) had median (range) CRP = 3 [0, 27] mg/dL, LDH = 484 [78, 3.745] U/I, and D-dimer = 1.133 [96, 55.660] ng/mL. The strongest risk factors were D-dimer, age, and cardiovascular disease. The model implemented using the variables identified using the LASSO Cox regression analysis classified 90% of non-survivors as high-risk individuals in the testing dataset. In this sample, the estimated median survival in the high-risk group was 9 days (95% CI; 9-37), while the low-risk group did not reach the median survival of 50% (p < 0.001). CONCLUSIONS: A machine learning model based on combined data available on the first days of hospitalization (demographics, CT-radiomics, comorbidities, and blood biomarkers), can identify SARS-CoV-2 patients at risk of serious illness and death.

Prognostic significance of peripheral consolidations at chest x-ray in severe COVID-19 pneumonia.

To evaluate the possible prognostic significance of the development of peripheral consolidations at chest x-ray in COVID-19 pneumonia, we retrospectively studied 92 patients with severe respiratory failure (PaO2/FiO2 ratio < 200 mmHg) that underwent at least two chest x-ray examinations (baseline and within 10 days of admission). Patients were divided in two groups based on the evolution of chest x-ray toward the appearance of peripheral consolidations or toward a greater extension of the lung abnormalities but without peripheral consolidations. Patients who developed lung abnormalities without peripheral consolidations as well as patients who developed peripheral consolidations showed, at follow-up, a significant worsening of the PaO2/FiO2 ratio but a significantly lower mortality and intubation rate was observed in patients with peripheral consolidations at chest x-ray. The progression of chest x-ray toward peripheral consolidations is an independent prognostic factor associated with lower intubation rate and mortality.

Clinical value of mesothelin in pleural effusions versus histology by medical thoracoscopy: brief report.

The aim of this study based on the third phase of the architecture of diagnostic research was to assess the sensitivity and specificity of soluble mesothelin-related peptide (SMRP) in pleural exudative effusions (PE) compared to the histology obtained by medical thoracoscopy as the diagnostic gold standard examination. We assessed 104 consecutive thoracoscopies. SMRP concentrations were obtained using an ELISA test. We had 34 mesotheliomas (25 epithelioid and 9 sarcomatoid), 35 pleural metastases, and 35 benign diseases. PE-SMRP were significantly higher in patients with epitheliomorphic mesothelioma (mean ± SD 46.55 ± 44.29 nM) than in patients with sarcomatoid mesothelioma (16.11 ± 25.02 nM; p = 0.061), pleural metastasis (7.52 ± 10.77 nM; p < 0.0001), or benign diseases (5.82 ± 8.86 nM; p < 0.0001). Using ROC curve analysis, PE-SMRP offered an AUC of 0.767 in its ability to differentiate between patients with mesothelioma and all other diagnoses. The diagnostic sensitivity and specificity of PE-SMRP for distinguishing mesothelioma from all other causes of pleural effusion, at a cut-off value of 19.6 nM, were 58.8 and 97.1 %, respectively. PE-SMRP higher than the assumed cut-off of 19.6 nM were observed in 18/25 (72.0%) epitheliomorphic mesotheliomas, 2/9 (22.2%) sarcomatoid mesotheliomas, 5/35 (14.3%) pleural metastases, and 1/35 (2.9%) benign diseases. We conclude that PE-SMRP adds some clinical information in the work-up of patients with a PE of unknown origin: (1) thoracoscopy should always be done in patients with a positive mesothelin; (2) a negative mesothelin does not exclude a malignant disease.

Evaluation of soluble mesothelin-related peptide as a diagnostic marker of malignant pleural mesothelioma effusions: its contribution to cytology.

Soluble mesothelin-related peptide (SMRP) is regarded as an FDA approved biomarker for the diagnosis and monitoring of pleural malignant mesothelioma (MPM). We detected the SMRP levels in pleural effusions (PE) by means of an ELISA and analyzed their diagnostic relevance to differentiate MPM from benign pathology and from non-MPM pleural metastasis. Comparison with cytology in MPM-PE was also performed. We found that SMRP detection in MPM-PE can help the diagnosis of MPM and provide additional diagnostic value to cytology. We concluded that SMRP test may be incorporated into clinical practice of PE from patients suspicious for MPM.

Clinical relevance of human mammaglobin mRNA in pleural effusion from patients undergoing thoracoscopy: a pilot study.

BACKGROUND: human mammaglobin (hMAM) expression has been reported in pleural effusions (PE). The aim of this study was to assess the clinical relevance of hMAM mRNA in PE from patients who underwent thoracoscopy.
 MATERIAL AND METHODS: A total of 288 patients with PE were studied, 155 of which were diagnosed with malignant and 133 with non-malignant diseases by thoracoscopy. Cells from PE were analyzed by nested hMAM RT-PCR. Statistical analyses were performed to evaluate the diagnostic performance parameters (DPP), the association between hMAM expression and benign or malignant status and the relative risk of cancer for patients with negative thoracoscopy showing hMAM positivity.
 RESULTS: hMAM mRNA was found in 68/288 (23.6%) PE samples of which 51 were from the 155 patients diagnosed with malignant diseases and 17 were from the 133 patients diagnosed with non-malignant diseases. A significant correlation between hMAM expression and malignancy was found (OR=3.04) and the DPP were as follows: sensitivity=32.9%, specificity=87.2%, accuracy=58.0%, positive predictive value=75.0% and negative predictive value=52.7%. Among the patients with negative thoracoscopy (n=133), 5/17 (29.4%) hMAM-positive patients had or developed a tumor during the 18-month follow up period, as compared to 10/116 (8.6%) hMAM-negative patients (relative risk of 4.6 for developing a malignancy).
 CONCLUSION: These findings suggest a possible application of hMAM RT-PCR detection in PE as to identify a false-negative thoracoscopy in non-specific pleuritis.

Diagnosis and origin determination of malignant pleural effusions through the use of the breast cancer marker human mammaglobin.

As was reported that human mammaglobin (hMAM) may be expressed in malignant pleural effusions (PEs), we investigated the relevance of hMAM reverse-transcriptase polymerase chain reaction (RT-PCR) for their diagnosis and determination of primary origin. Two hundred and twenty-eight malignant (132 male, 96 female) and 185 benign (132 male, 53 female) PEs were investigated. Statistical analyses evaluated the diagnostic performance parameters in all PEs and in cytologically negative malignant PEs, the association between hMAM and benign or malignant status by the direct index of correlation [diagnostic odds ratio (DOR)], chi test, and P value (P). In addition, the discriminative diagnostic power of hMAM expression, independently in breast cancer, lung cancer (LC), malignant mesothelioma (MM), and other cancers was evaluated. In the entire patient population, hMAM was detected in 45.6% and 5.4% of malignant and benign PEs, respectively, in the male group in 41.7% and 4.5% and in the female group in 51.0% and 7.5% of malignant and benign PEs, respectively. A statistically significant correlation between hMAM and malignancy was found in the entire population (DOR=14.68, P<0.001) and in the male (DOR=15.00, P<0.001) or female (DOR=12.77, P<0.001) groups. hMAM RT-PCR increased the diagnostic rate of malignant PEs as it allowed us to detect as malignant 32.1% of cytologically negative PEs. In female patients the positivity of hMAM indicated with higher probability (50.8%) the origin of PEs from breast cancer but lower probability from LC (17%), MM (9.4%), or other cancers (15.1%), whereas in male patients it indicated with similar probability (about 40%) the origin from LC or MM. Our results suggest that hMAM RT-PCR may provide information both in the diagnosis of PE and in the search for the primary site of neoplasia, either in male or female patients.