Differential privacy preserved federated learning for prognostic modeling in COVID-19 patients using large multi-institutional chest CT dataset.

BACKGROUND: Notwithstanding the encouraging results of previous studies reporting on the efficiency of deep learning (DL) in COVID-19 prognostication, clinical adoption of the developed methodology still needs to be improved. To overcome this limitation, we set out to predict the prognosis of a large multi-institutional cohort of patients with COVID-19 using a DL-based model. PURPOSE: This study aimed to evaluate the performance of deep privacy-preserving federated learning (DPFL) in predicting COVID-19 outcomes using chest CT images. METHODS: After applying inclusion and exclusion criteria, 3055 patients from 19 centers, including 1599 alive and 1456 deceased, were enrolled in this study. Data from all centers were split (randomly with stratification respective to each center and class) into a training/validation set (70%/10%) and a hold-out test set (20%). For the DL model, feature extraction was performed on 2D slices, and averaging was performed at the final layer to construct a 3D model for each scan. The DensNet model was used for feature extraction. The model was developed using centralized and FL approaches. For FL, we employed DPFL approaches. Membership inference attack was also evaluated in the FL strategy. For model evaluation, different metrics were reported in the hold-out test sets. In addition, models trained in two scenarios, centralized and FL, were compared using the DeLong test for statistical differences. RESULTS: The centralized model achieved an accuracy of 0.76, while the DPFL model had an accuracy of 0.75. Both the centralized and DPFL models achieved a specificity of 0.77. The centralized model achieved a sensitivity of 0.74, while the DPFL model had a sensitivity of 0.73. A mean AUC of 0.82 and 0.81 with 95% confidence intervals of (95% CI: 0.79-0.85) and (95% CI: 0.77-0.84) were achieved by the centralized model and the DPFL model, respectively. The DeLong test did not prove statistically significant differences between the two models (p-value = 0.98). The AUC values for the inference attacks fluctuate between 0.49 and 0.51, with an average of 0.50 ± 0.003 and 95% CI for the mean AUC of 0.500 to 0.501. CONCLUSION: The performance of the proposed model was comparable to centralized models while operating on large and heterogeneous multi-institutional datasets. In addition, the model was resistant to inference attacks, ensuring the privacy of shared data during the training process.

Hands and feet radiologic involvements in systemic sclerosis.

AIM: Systemic sclerosis (SSc) is a rare autoimmune disorder characterized by vascular and fibrosing involvement of the skin and internal organs. In this study, we determined the prevalence and characteristics of radiological hands and feet involvements in Iranian SSc patients to identify the associations between clinical features and radiologic findings. METHODS: 43 SSc patients (41 women and 2 men), with a median age of 44.8 years (ranges 26-70 years) and a mean disease duration of 11.8 years (ranges 2-28 years) were studied in this cross-sectional study. RESULTS: 42 patients had radiological changes both in their hands and feet. Only one patient had alteration just in hand. The most frequent changes that we found in hand were Juxta-articular Osteoporosis (93%), Acro-osteolysis (58.2%), and Joint Space Narrowing (55.8%). The prevalence of joint space narrowing or acro-osteolysis was higher in subjects with active skin involvement [modified Rodnan skin score (mRSS) > 14] [16/21 vs. 4/16 for patients with inactive skin involvement (mRSS < 14); p = 0.002]. The most frequent changes that we found in the foot were Juxta-articular Osteoporosis (93%), Acro-osteolysis (46.5%), Joint Space Narrowing (58.1%), and subluxation (44.2%). The presence of anti-ccp antibody was detected in 4 (9.3%), while positive rheumatoid factor was found in 13 (30.2%) of SSc patients. CONCLUSION: This study corroborates that arthropathy is common in SSc patients. The introduction of the specific radiological involvements of SSc needs to be confirmed by further studies, in order to define the appropriate prognosis and treatment of patients.

Primary Biliary Tract Cancers in Golestan, Iran: 13-Year Experience of Golestan Population-Based Cancer Registry.

BACKGROUND: Epidemiological research on the high-risk population might be helpful in early detection and prevention of biliary tract malignancies. This study assesses the prevalence of biliary tract cancer (BTC) in the Golestan province, northeastern Iran, between 2004 and 2016. METHODS: The current study used information from the Golestan Population-based Cancer Registry (GPCR) to access the epidemiology of BTC across a 13-year period while taking into account temporal and geographic differences. The number of cases, crude rates, age-standardized incidence rates (ASRs) per 100,000 person-years, average annual percent change (AAPC), age-specific incidence rates, and 95% confidence intervals (CI) were reported for each year with respect to gender and place of residence. RESULTS: Totally, 224 instances of BTC overall (54% of whom were females) were reported throughout the research period. The ASR of BTC was 1.7 (95% CI: 1.4‒2) for females and 1.4 (95% CI: 1.1‒1.6) for men, respectively. Males exhibited a growing time trend in incidence (AAPC: 7.18; CI: 0.06‒14.81; P-value:0.048), whereas females had a decreasing trend (AAPC: 0.82; CI: -5.94‒4.57; P-value: 0.740). Both sexes saw an increase in age-specific incidence rates starting at the age of 45; however, males experienced a significant increase in incidence in the age group of 75 to 79 while the female rates grew steadily. CONCLUSION: The focus for cancer control in this region may be given to demographic groups with a combination of risk factors, including male gender, older age, and urban residence.

COVID-19 prognostic modeling using CT radiomic features and machine learning algorithms: Analysis of a multi-institutional dataset of 14,339 patients.

BACKGROUND: We aimed to analyze the prognostic power of CT-based radiomics models using data of 14,339 COVID-19 patients. METHODS: Whole lung segmentations were performed automatically using a deep learning-based model to extract 107 intensity and texture radiomics features. We used four feature selection algorithms and seven classifiers. We evaluated the models using ten different splitting and cross-validation strategies, including non-harmonized and ComBat-harmonized datasets. The sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) were reported. RESULTS: In the test dataset (4,301) consisting of CT and/or RT-PCR positive cases, AUC, sensitivity, and specificity of 0.83 ± 0.01 (CI95%: 0.81-0.85), 0.81, and 0.72, respectively, were obtained by ANOVA feature selector + Random Forest (RF) classifier. Similar results were achieved in RT-PCR-only positive test sets (3,644). In ComBat harmonized dataset, Relief feature selector + RF classifier resulted in the highest performance of AUC, reaching 0.83 ± 0.01 (CI95%: 0.81-0.85), with a sensitivity and specificity of 0.77 and 0.74, respectively. ComBat harmonization did not depict statistically significant improvement compared to a non-harmonized dataset. In leave-one-center-out, the combination of ANOVA feature selector and RF classifier resulted in the highest performance. CONCLUSION: Lung CT radiomics features can be used for robust prognostic modeling of COVID-19. The predictive power of the proposed CT radiomics model is more reliable when using a large multicentric heterogeneous dataset, and may be used prospectively in clinical setting to manage COVID-19 patients.

Clinical Presentations of the Survivor and Non-survivor Hospitalized Patients with COVID-19 in the Golestan Province of Iran during the First Peak of the Epidemics.

Background: Considering the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) pandemic, which causes coronavirus disease 2019 (COVID-19), we aimed to report the clinical features of 427 patients with COVID-19 and the outcomes after one-month admission to major teaching hospitals in the northeast of Iran. Materials and Methods: Data of patients hospitalized with COVID-19 from 20 February 2020 to 20 April 2020 was analyzed using the R software. The cases and their outcomes were monitored up to one month following their admission. Results: Among 427 patients with a median age of 53 years (50.8% male), 81 (19%) were directly admitted to the ICU ward, and 68 (16%) died during the study. The mean (SD) lengths of hospital stay were significantly higher in the non-survivors (6 (9) days) than survivors (4 (5) days) (P = 0.018). Ventilation need was reported in 67.6% of the non-survivors and 0.8% of the survivors (P < 0.001). Cough (72.8%), fever (69.3%), and dyspnea (64.0%) were the most common symptoms. There were more comorbidities in the severe cases (73.5%) and non-survivor (77.5%). Liver and kidney damage were significantly more common in non-survivors. Ninety percent of the patients had at least one abnormal chest CT scan finding, including crazy paving and consolidation patterns (27.1%), followed by the ground-glass opacity (24.7%). Conclusion: Results showed that the patients' age, underlying comorbidities, levels of SpO2, and laboratory findings at the time of admission may predict the progress of the disease and can be considered mortality-related factors.

From bowel inflammation to the bone and joints: musculoskeletal examination in inflammatory bowel disease (IBD).

BACKGROUND: One of the most important complications in inflammatory Bowel Disease (IBD) are musculoskeletal manifestations that are reported in more than 50% of patients. OBJECTIVES: In this study, we aimed to evaluate the musculoskeletal and radiologic manifestations in our IBD patients. METHODS: In this cross-sectional study on 96 mild-to-moderate IBD patients (76 UC, 18 CD and 2 undifferentiated IBD) with mean (SD) age of 39.28 (11.42) years, 44 (45.8%) were males and 52 were (54.2%) females. Patients were examined by an expert rheumatologist and their musculoskeletal symptoms were assessed. The musculoskeletal system was evaluated by Modified Schober test, Thoracic expansion (TE), Occiput to wall distance (OWD), and Patrick's or FABER test. Peripheral joints were also examined in all four extremities. Then patients were referred for pelvic and lumbosacral x-ray. Sacroiliitis grading was performed using the New York criteria. RESULTS: Inflammatory low back pain was reported in 5 (5.2%), enthesopathy in 6 (6.5%) and dactylitis in 1 (1.1%). Positive Schober test was recorded in 5 (5.2%) and Patrick test in 3 (3.1%). Forty-nine (51%) cases had normal imaging with no sacroiliitis, endplate sclerosis was seen in 33 cases (34.4%), grade 3 and grade 4 were seen in 10 cases (10.4%). CONCLUSIONS: In the present study, 34.4% of the IBD patients had mild radiologic changes as endplate sclerosis and 95% had a normal physical examination.

Primary Liver Cancer in Golestan Province, Northeastern Iran: 13-Year Experience of Golestan Population-Based Cancer Registry (GPCR).

BACKGROUND: Liver cancer (LC) is among the most common and fatal cancers worldwide. A four-fold increasing trend is reported for LC age-standardized mortality rate (ASMR) in Iran within 1990-2015. In the present study, we aimed to report the incidence rate of LC during the 13-year establishment of the Golestan population-based cancer registry (GPCR). METHODS: Age-standardized incidence rate (ASIR) per 100000 person-years was calculated using CanReg5 software. The Joint point regression analysis was used to assess the temporal trends in incidence rate of LC. RESULTS: During the study period, GPCR registered 575 of LC cases, of which 373 (64.9%) were men and 297 (51.6%) lived in rural areas. The mean (SD) age of LC in men and women were 60.3 (17.2) and 58.0 (17.1), respectively. The ASIR of LC was significantly higher in men than women (4.7 vs. 2.6; P value<0.01) and a minimal increasing trend in LC incidence was observed in both sexes. The incidence rate of LC was significantly higher in rural areas compared to the urbans (3.91 vs. 3.40; P value=0.04). In the last 13 years, rural population had a significant increasing trend in ASIR of LC (average annual percent change [AAPC]=7.85, P value=0.005). CONCLUSION: LC was more prevalent in men among both urban and rural populations. The results showed a significant increasing trend in rural areas that requires Golestan health care system to take action in controlling the burden of LC in rural areas.

Can spontaneous pneumothorax be resolved in COVID-19 without hospital care? A case report.

BACKGROUND: We encountered the novel coronavirus infection as a pandemic in 2020. The infection started in Wuhan, China, and spread rapidly all over the world. CT scan has been used as an important diagnostic method in the detection of suspicious patients. One of the uncommon complications of coronavirus disease 2019 (COVID-19) is pneumothorax. CASE PRESENTATION: A 47-year-old smoker male with COVID-19 diagnosis, good general condition and no respiratory complaint, complicated by pneumothorax. He refused hospitalization. After educating him about the red flags and quarantine protocols, he continued treatment at home .Cap amoxicillin/clavulanic acid 625mg was prescribed for one week. A follow-up CT represented only small involvement of lungs. Pneumothorax was resolved spontaneously without any medical intervention and hospitalization. O2 saturation was in normal range an there was no dry cough anymore. CONCLUSION: According to our clinical experience, pneumothorax is resolved spontaneously in a COVID-19 case. Considering general status and hemodynamic stability, it is suggested to reduce invasive interventions in COVID-19 cases with pneumothorax.

Pneumothorax, Pneumomediastinum, and Subcutaneous Emphysema as Complications of COVID-19 Infection: A Case Series.

The novel coronavirus infection is a global health concern in 2020. Computerized tomography (CT) scan has an important role in diagnosis and follow-up with the course of the disease. The most common radiologic findings in patients are bilateral peripheral patchy ground-glass opacities and consolidations. Although in a few cases, as we reported, we encountered some rare manifestations such as pneumothorax, pneumomediastinum, and subcutaneous emphysema, which imply distinct concerns about the management and outcome of the disease. Pulmonary interstitial emphysema develops due to an increase in alveolar pressure or because of alveolar rupture, secondary to alveolar membrane damage by the virus and proceed to such a complication. Therefore, it is crucial to be aware of the complications of novel coronavirus infection in the deterioration of the disease.

Letter to the editor: efficacy of different methods of combination regimen administrations including dexamethasone, intravenous immunoglobulin, and interferon-beta to treat critically ill COVID-19 patients: a structured summary of a study protocol for a randomized controlled trial.

OBJECTIVES: There is little information about Coronavirus Disease 2019 (COVID-19) management for critically ill patients. Most of these patients develop acute respiratory distress syndrome (ARDS) due to excessive inflammatory response and the ensuing cytokine storm. Anti-inflammatory drugs including corticosteroids can be used to effectively reduce the effect of this cytokine storm and lung damage. However, corticosteroids can have side effects, so simultaneous administration of immunoglobulin (IV-IG) and interferon-beta can help manage treatment using corticosteroids. Therefore, we designed a trial to test our hypothesis that early administration of dexamethasone in combination with IV-IG and interferon-beta can reduce the effect of the cytokine storm in critically ill patients COVID-19. TRIAL DESIGN: A phase two multi-center randomized controlled trial (RCT) with three parallel arms (1:1:1 ratio). PARTICIPANTS: They will be hospitalized patients with severe COVID-19 who have positive RT-PCR test and have blood oxygen saturation levels (SpO2) less than 90% and respiratory rate higher than 24 per minute or have involvement of more than 50% of their lung when viewed using computed tomography (CT)-scan. The age range of patients will be 18-70 years old. EXCLUSION CRITERIA: the need for intubation; allergy, intolerance, or contraindication to any study drug including dexamethasone, IV-IG, and interferon-beta; pregnancy or lactation; known HIV positive or active hepatitis B or C. The study will be conducted in several hospitals of the Golestan province, Iran. INTERVENTION AND COMPARATOR: The study subjects will be randomly allocated to three treatment arms: two experimental groups (two arms: Intervention 1 and Intervention 2) and one Control Group, which will be matched for age and sex using frequency matching method. Each eligible patient in the control arm will receive the standard treatment for COVID-19 based on WHO guidelines and the Ministry of the Health and Medical Education (MOHME) of Iran. Each patient in the Intervention Group 1 will receive the standard treatment for COVID-19 and dexamethasone, at the first 24 hours' time of admission. The intervention begins with the administration of dexamethasone based on the SpO2 levels. If the level of SpO2 does not improve after 24 hours, IV-IG (400 mg/kg once daily for 5 days) and interferon-beta (7 doses every other day) will be prescribed along with dexamethasone administration. In Intervention Group 2, the administration of dexamethasone will be started within the first 24 hours' time of admission and will be continued for 48-72 hours and then the SpO2 level will be checked. Then, if the level of SpO2 has not improved after that time, IV-IG and interferon-beta will be prescribed as the same dosage as Group 1. If the percentages of the SpO2 level are between 85 and 90/ 80 and 85/ 75 and 80/ less than 75, the dosages will be 4 mg every 12 hours/ 4 mg every 8 hours/ 8 mg every 12 hours/ 8 mg every 8 hours, respectively. According to the WHO recommendation, all participants will have the best available supportive care with full monitoring. MAIN OUTCOMES: Primary: An increase in the SpO2 level to reach more than 90% in each case, which will be assessed by the oximeter. Secondary: The duration of hospital stays; intubation status and the percentage of patients who are free of mechanical ventilation; the mortality rates during hospitalization and one month after the admission time. RANDOMISATION: Participants will be allocated into either control or intervention groups with a 1:1:1 allocation ratio using a computer random number generator to generate a table of random numbers for simple randomization. BLINDING (MASKING): The project's principal investigator (PI) is unblinded. However, the PI will not analyse the data and interpret the results. An unblinded researcher (a pharmacist) will cover the drug's bottles with aluminium foil and prepare them interventions and control drugs in a syringe with a code so that patients are blinded. This person will have no patients contact. The staff and nurses, caring for the patients, will be unblinded for each study group due to the nature of this study. The staff that take outcome measurements will be blinded. The laboratory technicians will also be blinded as well as the statistical team. These study statisticians will have access to coded data and will analyse the data labelled as group X, group Y, and group Z. NUMBERS TO BE RANDOMISED (SAMPLE SIZE): The target sample size will be 105 critically ill COVID-19 patients, who will be allocated randomly to the three trial arms with 35 patients in each group. TRIAL STATUS: Recruitment is ongoing. The study began on April 18 2020 and will be completed June 19 2020. This summary describes protocol version 1; April 2 2020. TRIAL REGISTRATION: https://www.irct.ir/. Identifier: IRCT20120225009124N4 version 1; Registration date: April 2 2020. FULL PROTOCOL: The full protocol is attached as an additional file, accessible from the Trials website (Additional file 1). In the interest in expediting the dissemination of this material, the familiar formatting has been eliminated; this Letter serves as a summary of the key elements of the full protocol. The full protocol has been reported in accordance with the Standard Protocol Items: Recommendations for Clinical Interventional Trials (SPIRIT) guidelines.