Deep learning-based overall survival prediction model in patients with rare cancer: a case study for primary central nervous system lymphoma.

PURPOSE: Primary central nervous system lymphoma (PCNSL) is a rare, aggressive form of extranodal non-Hodgkin lymphoma. To predict the overall survival (OS) in advance is of utmost importance as it has the potential to aid clinical decision-making. Though radiomics-based machine learning (ML) has demonstrated the promising performance in PCNSL, it demands large amounts of manual feature extraction efforts from magnetic resonance images beforehand. deep learning (DL) overcomes this limitation. METHODS: In this paper, we tailored the 3D ResNet to predict the OS of patients with PCNSL. To overcome the limitation of data sparsity, we introduced data augmentation and transfer learning, and we evaluated the results using r stratified k-fold cross-validation. To explain the results of our model, gradient-weighted class activation mapping was applied. RESULTS: We obtained the best performance (the standard error) on post-contrast T1-weighted (T1Gd)-area under curve [Formula: see text], accuracy [Formula: see text], precision [Formula: see text], recall [Formula: see text] and F1-score [Formula: see text], while compared with ML-based models on clinical data and radiomics data, respectively, further confirming the stability of our model. Also, we observed that PCNSL is a whole-brain disease and in the cases where the OS is less than 1 year, it is more difficult to distinguish the tumor boundary from the normal part of the brain, which is consistent with the clinical outcome. CONCLUSIONS: All these findings indicate that T1Gd can improve prognosis predictions of patients with PCNSL. To the best of our knowledge, this is the first time to use DL to explain model patterns in OS classification of patients with PCNSL. Future work would involve collecting more data of patients with PCNSL, or additional retrospective studies on different patient populations with rare diseases, to further promote the clinical role of our model.

Radiomics-Based Machine Learning Model for Predicting Overall and Progression-Free Survival in Rare Cancer: A Case Study for Primary CNS Lymphoma Patients.

Primary Central Nervous System Lymphoma (PCNSL) is an aggressive neoplasm with a poor prognosis. Although therapeutic progresses have significantly improved Overall Survival (OS), a number of patients do not respond to HD-MTX-based chemotherapy (15-25%) or experience relapse (25-50%) after an initial response. The reasons underlying this poor response to therapy are unknown. Thus, there is an urgent need to develop improved predictive models for PCNSL. In this study, we investigated whether radiomics features can improve outcome prediction in patients with PCNSL. A total of 80 patients diagnosed with PCNSL were enrolled. A patient sub-group, with complete Magnetic Resonance Imaging (MRI) series, were selected for the stratification analysis. Following radiomics feature extraction and selection, different Machine Learning (ML) models were tested for OS and Progression-free Survival (PFS) prediction. To assess the stability of the selected features, images from 23 patients scanned at three different time points were used to compute the Interclass Correlation Coefficient (ICC) and to evaluate the reproducibility of each feature for both original and normalized images. Features extracted from Z-score normalized images were significantly more stable than those extracted from non-normalized images with an improvement of about 38% on average (p-value < 10-12). The area under the ROC curve (AUC) showed that radiomics-based prediction overcame prediction based on current clinical prognostic factors with an improvement of 23% for OS and 50% for PFS, respectively. These results indicate that radiomics features extracted from normalized MR images can improve prognosis stratification of PCNSL patients and pave the way for further study on its potential role to drive treatment choice.

Assessing the added value of apparent diffusion coefficient, cerebral blood volume, and radiomic magnetic resonance features for differentiation of pseudoprogression versus true tumor progression in patients with glioblastoma.

Background: Pseudoprogression (PsPD) is a major diagnostic challenge in the follow-up of patients with glioblastoma (GB) after chemoradiotherapy (CRT). Conventional imaging signs and parameters derived from diffusion and perfusion-MRI have yet to prove their reliability in clinical practice for an accurate differential diagnosis. Here, we tested these parameters and combined them with radiomic features (RFs), clinical data, and MGMT promoter methylation status using machine- and deep-learning (DL) models to distinguish PsPD from Progressive disease. Methods: In a single-center analysis, 105 patients with GB who developed a suspected imaging PsPD in the first 7 months after standard CRT were identified retrospectively. Imaging data included standard MRI anatomical sequences, apparent diffusion coefficient (ADC), and normalized relative cerebral blood volume (nrCBV) maps. Median values (ADC, nrCBV) and RFs (all sequences) were calculated from DL-based tumor segmentations. Generalized linear models with LASSO feature-selection and DL models were built integrating clinical data, MGMT methylation status, median ADC and nrCBV values and RFs. Results: A model based on clinical data and MGMT methylation status yielded an areas under the receiver operating characteristic curve (AUC) = 0.69 (95% CI 0.55-0.83) for detecting PsPD, and the addition of median ADC and nrCBV values resulted in a nonsignificant increase in performance (AUC = 0.71, 95% CI 0.57-0.85, P = .416). Combining clinical/MGMT information with RFs derived from ADC, nrCBV, and from all available sequences both resulted in significantly (both P < .005) lower model performances, with AUC = 0.52 (0.38-0.66) and AUC = 0.54 (0.40-0.68), respectively. DL imaging models resulted in AUCs ≤ 0.56. Conclusion: Currently available imaging biomarkers could not reliably differentiate PsPD from true tumor progression in patients with glioblastoma; larger collaborative efforts are needed to build more reliable models.

Continuous sensing of IFNα by hepatic endothelial cells shapes a vascular antimetastatic barrier.

Hepatic metastases are a poor prognostic factor of colorectal carcinoma (CRC) and new strategies to reduce the risk of liver CRC colonization are highly needed. Herein, we used mouse models of hepatic metastatization to demonstrate that the continuous infusion of therapeutic doses of interferon-alpha (IFNα) controls CRC invasion by acting on hepatic endothelial cells (HECs). Mechanistically, IFNα promoted the development of a vascular antimetastatic niche characterized by liver sinusoidal endothelial cells (LSECs) defenestration extracellular matrix and glycocalyx deposition, thus strengthening the liver vascular barrier impairing CRC trans-sinusoidal migration, without requiring a direct action on tumor cells, hepatic stellate cells, hepatocytes, or liver dendritic cells (DCs), Kupffer cells (KCs) and liver capsular macrophages (LCMs). Moreover, IFNα endowed LSECs with efficient cross-priming potential that, along with the early intravascular tumor burden reduction, supported the generation of antitumor CD8+ T cells and ultimately led to the establishment of a protective long-term memory T cell response. These findings provide a rationale for the use of continuous IFNα therapy in perioperative settings to reduce CRC metastatic spreading to the liver.

Colorectal cancer remains one of the most widespread and deadly cancers worldwide. Poor health outcomes are usually linked to diseased cells spreading from the intestine to create new tumors in the liver or other parts of the body. Treatment involves surgically removing the initial tumors in the bowel, but patient survival could be improved if, in parallel, their immune system was 'boosted' to destroy cancer cells before they can form other tumors. Interferon alpha is a small protein which helps to coordinate how the immune system recognizes and deactivates foreign agents and cancerous cells. It has recently been trialed as a colorectal cancer treatment to prevent tumors from spreading to the liver, but only with limited success. This partly because interferon-alpha is usually administered in high and pulsed doses, which cause severe side effects through the body. Instead, Tran, Ferreira, Alvarez-Moya et al. aimed to investigate whether continuously delivering lower amounts of the drug could be a better approach. This strategy was tested on mice in which colorectal cancer cells had been implanted into the wall of the large intestine. Continuous administration minimized the risk of the implanted cancer cells spreading to the liver while also creating fewer side effects. The team was able to identify an optimum delivery strategy by varying how much interferon-alpha the animals received and when. Further experiments also revealed a new mechanism by which interferon-alpha prevented the spread of colorectal cancer. Upon receiving continuous doses of the drug, a group of liver cells started to generate a physical barrier which stopped cancer cells from being able to invade the organ. The treatment also promoted long-term immune responses that targeted diseased cells while being safe for healthy tissues. If confirmed in clinical trials, these results suggest that colorectal patients undergoing tumor removal surgery may benefit from also receiving interferon-alpha through continuous delivery.

Impact of clinical and subclinical coronary artery disease as assessed by coronary artery calcium in COVID-19.

BACKGROUND AND AIMS: The potential impact of coronary atherosclerosis, as detected by coronary artery calcium, on clinical outcomes in COVID-19 patients remains unsettled. We aimed to evaluate the prognostic impact of clinical and subclinical coronary artery disease (CAD), as assessed by coronary artery calcium score (CAC), in a large, unselected population of hospitalized COVID-19 patients undergoing non-gated chest computed tomography (CT) for clinical practice. METHODS: SARS-CoV 2 positive patients from the multicenter (16 Italian hospitals), retrospective observational SCORE COVID-19 (calcium score for COVID-19 Risk Evaluation) registry were stratified in three groups: (a) "clinical CAD" (prior revascularization history), (b) "subclinical CAD" (CAC >0), (c) "No CAD" (CAC = 0). Primary endpoint was in-hospital mortality and the secondary endpoint was a composite of myocardial infarction and cerebrovascular accident (MI/CVA). RESULTS: Amongst 1625 patients (male 67.2%, median age 69 [interquartile range 58-77] years), 31%, 57.8% and 11.1% had no, subclinical and clinical CAD, respectively. Increasing rates of in-hospital mortality (11.3% vs. 27.3% vs. 39.8%, p < 0.001) and MI/CVA events (2.3% vs. 3.8% vs. 11.9%, p < 0.001) were observed for patients with no CAD vs. subclinical CAD vs clinical CAD, respectively. The association with in-hospital mortality was independent of in-study outcome predictors (age, peripheral artery disease, active cancer, hemoglobin, C-reactive protein, LDH, aerated lung volume): subclinical CAD vs. No CAD: adjusted hazard ratio (adj-HR) 2.86 (95% confidence interval [CI] 1.14-7.17, p=0.025); clinical CAD vs. No CAD: adj-HR 3.74 (95% CI 1.21-11.60, p=0.022). Among patients with subclinical CAD, increasing CAC burden was associated with higher rates of in-hospital mortality (20.5% vs. 27.9% vs. 38.7% for patients with CAC score thresholds≤100, 101-400 and > 400, respectively, p < 0.001). The adj-HR per 50 points increase in CAC score 1.007 (95%CI 1.001-1.013, p=0.016). Cardiovascular risk factors were not independent predictors of in-hospital mortality when CAD presence and extent were taken into account. CONCLUSIONS: The presence and extent of CAD are associated with in-hospital mortality and MI/CVA among hospitalized patients with COVID-19 disease and they appear to be a better prognostic gauge as compared to a clinical cardiovascular risk assessment.

Leakage Risk Stratification After Laparoscopic Sleeve Gastrectomy (LSG): Is There a Role for Routine Postoperative CT Scan?

PURPOSE: Leakage of the gastric remnant after laparoscopic sleeve gastrectomy (LSG) represents an unpredictable, dreadful occurrence. Our aim was to assess whether routine postoperative CT scan is an effective tool for early prediction of leakage after LSG. MATERIALS AND METHODS: From a prospectively acquired database, all consecutive patients who underwent LSG between January 2015 and December 2018 were identified; within this database, all patients who were evaluated with at least one contrast-enhanced CT scan within 48 h from surgery were enrolled in this retrospective study. The selected CT findings included twisting of the gastric remnant, perigastric air bubbles, and hematoma; the antral segment proximal from the pylorus to the first staple firing was also analyzed in terms of distance (StP, stapler to pylorus distance) and linearity (LI, linearity index). RESULTS: After exclusions, 250 patients were included; 10 patients suffered from gastric leakage. Patients with perigastric hematoma and/or twisting of the distal part of the gastric remnant on routine postoperative CT scan were found to be more likely to develop leakage after LSG (p = 0.005 and p < 0.001, respectively). The mean StP was 45 ± 19.1 mm; the mean LI was 1.54 ± 0.4. Patients with subsequent development of leakage had significantly lower StP (26.7 ± 12.5 mm vs. 45.9 ± 18.9 mm; p = 0.001) and LI values (1.16 ± 0.11 vs. 1.55 ± 0.39; p = 0.002). CONCLUSION: Routine postoperative CT scan after LSG permits early stratification of leakage risk, thus providing an actual aid for patients' management.