Graph-Based Fusion of Imaging and Non-Imaging Data for Disease Trajectory Prediction

This study proposes a graph convolutional neural networks (GCN) architecture for fusion of radiological imaging and non-imaging tabular electronic health records (EHR) for the purpose of clinical event prediction. We focused on a cohort of hospitalized patients with positive RT-PCR test for COVID-19 and developed GCN based models to predict three dependent clinical events (discharge from hospital, admission into ICU, and mortality) using demographics, billing codes for procedures and diagnoses and chest X-rays. We hypothesized that the two-fold learning opportunity provided by the GCN is ideal for fusion of imaging information and tabular data as node and edge features, respectively. Our experiments indicate the validity of our hypothesis where GCN based predictive models outperform single modality and traditional fusion models. We compared the proposed models against two variations of imaging-based models, including DenseNet-121 architecture with learnable classification layers and Random Forest classifiers using disease severity score estimated by pre-trained convolutional neural network. GCN based model outperforms both imaging-only methods. We also validated our models on an external dataset where GCN showed valuable generalization capabilities. We noticed that edge-formation function can be adapted even after training the GCN model without limiting application scope of the model. Our models take advantage of this fact for generalization to external data. © 2023 IEEE.

QUASAR Induction Study 1 Cumulative Response to Guselkumab in Patients With Moderately to Severely Active Ulcerative Colitis

Introduction: Guselkumab (GUS), an IL-23p19 antagonist, had greater efficacy than placebo (PBO) in achieving clinical response and clinical remission atWeek (Wk) 12 in the randomized, controlled Phase 2b QUASAR Induction Study 1 (NCT04033445) in patients with moderately to severely active ulcerative colitis (UC).1 Patients who were not in clinical response at Wk 12 received GUS treatment through Wk 24. Here, we report GUS cumulative efficacy and safety results for Induction Study 1. Method(s): Eligible patients had moderately to severely active UC (modified Mayo score of 5 to 9 with a Mayo endoscopy subscore >=2) at baseline. Patients were randomized 1:1:1 to IV GUS 200mg, 400mg, or PBO at Wks 0, 4, and 8. Patients who were not in clinical response to IV induction at Wk 12 received GUS treatment (PBO IV->GUS 200mg IV;GUS 200mg IV->GUS 200mg SC;GUS 400mg IV->GUS 200mg SC) at Wks 12, 16, and 20 and were evaluated at Wk 24 (Figure). Matching IV or SC PBO was administered to maintain the blind. Result(s): Three hundred thirteen patients were randomized and treated at baseline. Demographic and disease characteristics at baseline were similar among the treatment groups, and approximately 50% had a prior inadequate response or intolerance to advanced UC therapy. AtWk 12, clinical response was achieved by 61.4% (62/101) and 60.7% (65/107) of patients randomized to GUS 200mg and GUS 400mg IV vs 27.6 % (29/105) of patients randomized to PBO IV (both p< 0.001). Of the patients in the GUS groups who were not in clinical response at Wk 12, 54.3% (19/35) in the GUS 200mg IV->200mg SC group and 50.0% (19/38) in the GUS 400mg IV->200mg SC group achieved clinical response at Wk 24. Clinical response atWk 12 or 24 was achieved by 80.2% of patients who were randomized to GUS 200mg IV and 78.5% of patients who were randomized to GUS 400mg IV. For patients who received PBO IV->GUS 200mg IV, clinical response at Wk 24 (65.2%) was similar toWk 12 clinical response following GUS 200mg IV induction (61.4%). The most frequent adverse events among all GUS-treated pts (n=274) were anemia (7.7%), headache (5.1%), worsening UC (4.4%), COVID-19 (3.6%), arthralgia (2.9%) and abdominal pain (2.6%) which are consistent with Wk 12 results. Conclusion(s): Overall, approximately 80% of patients randomized to receive GUS achieved clinical response at Wk 12 or 24. Continued treatment with SC GUS allowed 50-54.3% of IV GUS Wk 12 clinical nonresponders to achieve clinical response at Wk 24. No new safety concerns for GUS were identified. (Figure Presented).

QUASAR Induction Study 1 Cumulative Response to Guselkumab in Patients With Moderately to Severely Active Ulcerative Colitis

Background: Guselkumab (GUS), an IL-23p19 antagonist, had greater efficacy than placebo (PBO) in achieving clinical response and clinical remission at Week (Wk) 12 in the randomized, controlled Phase 2b QUASAR Induction Study 1 (NCT04033445) in patients with moderately to severely active ulcerative colitis (UC).1 Patients who were not in clinical response at Wk 12 received GUS treatment through Wk 24. Here, we report GUS cumulative efficacy and safety results for Induction Study 1. Method(s): Eligible patients had moderately to severely active UC (modified Mayo score of 5 to 9 with a Mayo endoscopy subscore >=2) at baseline. Patients were randomized 1:1:1 to IV GUS 200mg, 400mg, or PBO at Wks 0, 4, and 8. Patients who were not in clinical response to IV induction at Wk 12 received GUS treatment (PBO IVGUS 200mg IV;GUS 200mg IV->GUS 200mg SC;GUS 400mg IV->GUS 200mg SC) at Wks 12, 16, and 20 and were evaluated at Wk 24 (Figure 1). Matching IV or SC PBO was administered to maintain the blind. Result(s): Three hundred thirteen patients were randomized and treated at baseline. Demographic and disease characteristics at baseline were similar among the treatment groups, and approximately 50% had a prior inadequate response or intolerance to advanced UC therapy. At Wk 12, clinical response was achieved by 61.4% (62/101) and 60.7% (65/107) of patients randomized to GUS 200mg and GUS 400mg IV vs 27.6% (29/105) of patients randomized to PBO IV (both p<0.001). Of the patients in the GUS groups who were not in clinical response at Wk 12, 54.3% (19/35) in the GUS 200mg IV->200mg SC group and 50.0% (19/38) in the GUS 400mg IV->200mg SC group achieved clinical response at Wk 24. Clinical response at Wk 12 or 24 was achieved by 80.2% of patients who were randomized to GUS 200mg IV and 78.5% of patients who were randomized to GUS 400mg IV. For patients who received PBO IV->GUS 200mg IV, clinical response at Wk 24 (65.2%) was similar to Wk 12 clinical response following GUS 200mg IV induction (61.4%). The most frequent adverse events among all GUS-treated pts (n=274) were anemia (7.7%), headache (5.1%), worsening UC (4.4%), COVID-19 (3.6%), arthralgia (2.9%) and abdominal pain (2.6%) which are consistent with Wk 12 results. Conclusion(s): Overall, approximately 80% of patients randomized to receive GUS achieved clinical response at Wk 12 or 24. Continued treatment with SC GUS allowed 50-54.3% of IV GUS Wk 12 clinical nonresponders to achieve clinical response at Wk 24. No new safety concerns for GUS were identified.

Treatment outcome of tofacitinib dose reduction to 5 mg BID vs remaining on 10 mg BID in patients with Ulcerative Colitis who were in stable remission on 10 mg BID: Updated 30-month data from the double-blind, randomised RIVETING study

Background: Tofacitinib is an oral small molecule Janus kinase inhibitor for the treatment of ulcerative colitis. The long-term, Phase 3b/4 RIVETING study (NCT03281304) assessed the efficacy and safety of tofacitinib dose reduction from tofacitinib 10 mg twice daily (BID) maintenance therapy to 5 mg BID in patients (pts) in stable remission. 1 We present a final analysis of the RIVETING study after >=30 months of treatment. Method(s): RIVETING was a double-blind, randomised, parallelgroup study with a 42-month (M) duration. Eligible pts had received tofacitinib 10 mg BID for >=2 consecutive years in an open-label, long-term extension study (NCT01470612), had been in stable remission for >=6 months and corticosteroid-free for >=4 weeks prior to baseline. The primary efficacy endpoint was remission at M6 based on modified Mayo (mMayo) score (endoscopic and stool frequency subscores of <=1 and rectal bleeding subscore of 0).1 RIVETING was terminated (primary objective met) when all pts had passed M30 (or discontinued);some pts had already completed all visits up to M42. Here, we assess efficacy at M30 and safety throughout. Result(s): Overall, 140 pts were randomised (1:1) to tofacitinib 5 or 10 mg BID;50.0% and 62.9% were in remission based on mMayo score at M30 in the 5 and 10 mg BID dose groups, respectively, with consistent findings observed for other secondary efficacy endpoints (Table 1). At M30, observed differences for mMayo remission between tofacitinib 10 and 5 mg BID were generally greater in the subgroup with a baseline endoscopic subscore of 1 vs 0 and in the subgroup with vs without prior tumour necrosis factor inhibitor (TNFi) failure (Table 1). The percentage of pts who experienced loss of remission by M30, as estimated from Kaplan-Meier curves, was numerically higher in the 5 vs the 10 mg BID dose group (28.1%;95% confidence interval [CI], 17.68-39.49 vs 21.7%;95% CI, 12.21-32.95, respectively). Table 2 shows adverse events of special interest, by dose group. One death occurred due to fatal coronavirus disease 2019 pneumonia (10 mg BID). Conclusion(s): These long-term data showed that most pts in stable remission on tofacitinib 10 mg BID maintenance therapy maintained mMayo score remission through M30 after dose reduction to 5 mg BID. Dose group difference in remission at M30 was consistent with that at M6.1 Differences in remission between the tofacitinib 5 and 10 mg BID groups were greater in subgroups with an endoscopic subscore of 1 vs 0 and in subgroups with vs without prior TNFi failure. Overall, safety findings were consistent with tofacitinib's known safety profile;incidence of serious infections and herpes zoster was numerically higher in the tofacitinib 10 vs 5 mg BID group. (Table Presented).

Trans-continental analysis of over, 2000 Inflammatory Bowel Disease patients implicates geography, disease type, and exposure to immunosuppression as drivers of SARS-CoV-2 seroprevalence: data from the ICARUS-IBD Consortium

Background IBD patients on immune-modulatory therapies are considered high-risk for SARS-CoV-2 infection. Direct comparisons of serological responses to SARS-CoV-2 infection in IBD patients across different continents and medications are lacking. We performed SARS-CoV-2 sero-surveillance of IBD patients prior to vaccination at seven large tertiary centres in Asia, Europe, and North America. Methods Clinical data and sera were collected from, 2,213 IBD patients receiving routine care at institutions in Belgium, Canada, Hong Kong, India, Japan, the United Kingdom, and the United States between, 26 May, 2020 and, 24 September, 2021 (Table, 1). Sera were taken prior to vaccination. Clinical data were collected through patient questionnaires and medical records. Antibody reactivity to the SARS-CoV-2 spike protein was assessed using the Roche SARS-CoV-2 anti-spike total antibody and/or Siemens Healthineers COV2T anti-spike total antibody assays, which showed, 99.4% concordance. Univariate analysis was performed to evaluate association between individual variables and sero-status. Results The pre-vaccination seroprevalence of antibodies to SARS-CoV-2 in IBD patient varied widely according to location from, 0% in Hong Kong to, 57.9% in New Delhi, India (p<0.001). Rates in Europe and North America were similar (range, 3.57%-8.94%). Overall, SARS-CoV-2 seroprevalence appears to be equal to or less than local populations (Table, 2). Seroprevalence rates were associated with IBD type (7.8% CD, 12.4% UC, 15% IBD-U, p<0.001), smoking status (p<0.001), and history of COVID diagnosis (p<0.001) or COVID hospitalization (p=0.001), and any IMM (p<0.001). (Table, 3). Whilst there were no significant differences in seroprevalence between patients receiving infliximab (IFX), vedolizumab (VDZ), and ustekinumab (UST), antibody levels were attenuated in patients on IFX monotherapy and combination therapy (both p=0.002) and VDZ (p=0.02), compared with no medications (Figure 1). Conclusion We confirm in diverse poulations that exposure to biologics or immunomodulators, type of disease, and smoking status are associated with seroprevalence and antibody levels. We show for the first time the dominant influence of geographical location on sero-status in these patients. These observations should be considered as we look towards post-vaccination data to help stratify patients for clinical guidelines on SARS-CoV-2 vaccination.

TRANS-CONTINENTAL ANALYSIS OF OVER 2,000 INFLAMMATORY BOWEL DISEASE PATIENTS IMPLICATES GEOGRAPHY, DISEASE TYPE, AND EXPOSURE TO IMMUNOSUPPRESSION AS DRIVERS OF SARS-COV-2 SEROPREVALENCE

Background: IBD patients on immune-modulatory therapies are considered high-risk for SARS-CoV-2 infection. Direct comparisons of serological responses to SARS-CoV-2 infection in IBD patients across different continents and medications are lacking. We performed SARSCoV- 2 sero-surveillance of IBD patients prior to vaccination at seven large tertiary centres in Asia, Europe, and North America. Methods: Clinical data and sera were collected from 2,241 IBD patients receiving routine care at institutions in Belgium, Canada, Hong Kong, India, Japan, United Kingdom, and the United States between May 2020 and September 2021 (Table 1). Sera were taken prior to vaccination. Clinical data were collected from patient questionnaires and medical records. Antibody reactivity to the SARS-CoV-2 spike protein was assessed using the Roche SARS-CoV-2 anti-spike total antibody and/or Siemens Healthineers COV2T anti-spike total antibody assays, which showed 99.4% concordance. We performed univariate analysis to evaluate association between variables and sero-status. Results: The pre-vaccination seroprevalence of antibodies to SARS-CoV-2 in IBD patient varied widely according to location from 0% in Hong Kong, China, to 57.9% in New Delhi, India. Rates in Europe and North America were similar (range 3.6%-8.9%). Overall, SARSCoV- 2 seroprevalence appears to be equal to or less than local populations (Table 1). Seroprevalence rates were associated with IBD type (Crohn's disease 7.8%, ulcerative colitis 12.4%, IBD-unclassified 15.0%, p<0.001), smoking status (p<0.001), and history of COVID diagnosis (p<0.001) or COVID hospitalization (p=0.001), and any immunomodulator (IMM) (p<0.001) (Table 1). Infection as indicated by seropositivity in the absence of known COVID infection occurred in 7.3% of patients. Whilst there were no significant differences in seroprevalence between patients receiving anti-tumor necrosis factor (anti-TNF) medications, vedolizumab (VDZ), and ustekinumab (UST), antibody levels were attenuated in patients on anti-TNF monotherapy (p=0.002), anti-TNF + IMM combination therapy (p=0.002), and VDZ (p=0.02), compared with no medications (Figure 1). Conclusion: We confirm in diverse populations that exposure to anti-TNFs, vedolizumab and immunomodulators, type of disease, and smoking status are associated with seroprevalence and antibody levels. We show for the first time the dominant influence of geographical location on sero-status in these patients. These observations should be considered as we look towards post-vaccination data to help stratify patients for clinical guidelines on SARS-CoV-2 vaccination. (Table Presented) Table 1. Seroprevalence of total anti-SARS-CoV-2 spike antibodies in IBD patients by ICARUS centre with non-IBD controls noted for New Delhi, India, and publicly reported local seroprevalence and by selected patient characteristics.(Figure Presented) Figure 1. Antibody levels by medication group.

POST-VACCINATION SYMPTOMS AFTER A THIRD DOSE OF SARS-COV-2 MRNA VACCINATION IN PATIENTS WITH INFLAMMATORY BOWEL DISEASE

Background:Symptoms after SARS-CoV-2 primary vaccination among patients with inflammatory bowel disease (IBD) are generally similar to the general population,although symptoms after the second dose are more frequent and severe than after the first dose.Postvaccination symptoms after a 3rd mRNA vaccine dose in the predominantly immune-compromised IBD population is unknown.Methods:Adults with IBD participating in the prospective Coronavirus Risk Associations and Longitudinal Evaluation in IBD (CORALE-IBD) vaccine registry who received a 3rd mRNA vaccine dose were asked to complete a detailed symptom survey 1 week after vaccination.Symptoms were assessed across 11 organ systems,and graded as mild,moderate,or severe,or requiring hospitalization.“Severe+” referred to those with severe symptoms or who required hospitalization.We stratified by age (<or> 50 years) given prior distinct symptom profiles after dose 2 (D2).We also evaluated whether severe+ symptoms after D2 predicted severe+ symptoms after dose 3 (D3).Results:We included 524 participants (70% female, mean age 45 years) who received a 3rd mRNA vaccine through October 11, 2021.Most had Crohn's disease (71%), and 89% were on biologic therapies.Most (58%) had received primary vaccination with BNT562b2,and only 3.5% reported prior COVID infection at the time of initial vaccination.Overall, 97% of subjects received a 3rd dose with the same mRNA vaccine as in their initial series with the remainder receiving the other mRNA vaccine type.No participants received a 3rd dose with the Ad26.CoV.2 (J&J) vaccine. Overall, 41% reported symptoms after a 3rd dose,with symptoms generally more frequent and severe among those <55 years (Table).The most frequent postvaccination symptom was injection site pain (39%).Common systemic symptoms included fatigue/malaise (34%),headache (23%),and muscle, bone or joint symptoms (13%).These were all less frequent after D3 than after D2 (Figure).Gastrointestinal symptoms were reported by 8.8%, which was slightly more frequent than after D2 (7.8%).Among those with postvaccination symptoms, the proportion with severe symptoms after D3 was lower than D2 for fatigue/ malaise, headache, dizziness and lightheadedness, fever/chills, and rheumatologic symptoms, but was slightly higher than D2 for gastrointestinal symptoms.Severe+ symptoms were seen in 17% after D2 and in 14% after D3. Of those with severe+ symptoms after D2, 34% had severe+ symptoms after D3.In contrast, about 22% had severe+ symptoms after D3 but did not report severe+ symptoms after D2.Conclusion:The frequency and severity of symptoms after a 3rd mRNA vaccine dose are generally similar or lower than those after a second dose.Furthermore, prior severe+ symptoms after D2 do not necessarily predict severe+ symptoms after D3. Further evaluation of postvaccination gastrointestinal symptoms in this population is warranted. (Figure Presented) (Table Presented)

PATIENTS WITH INFLAMMATORY BOWEL DISEASE HAVE IMPAIRED HUMORAL BUT PRESERVED CELLULAR RESPONSES TO SARS-COV-2 MRNA VACCINATION

Background: Vaccine-induced protection against SARS-CoV-2 infection is predominantly mediated by humoral immunity;protection against disease progression is primarily determined by cellular immunity. Patients with inflammatory bowel disease (IBD) have high rates of post-vaccination anti-Spike IgG [IgG(S)] seroconversion, but postvaccination immune responses relative to non-IBD controls have not been well described. We aimed to assess post-vaccination humoral (antibody) and cellular (T-cell) responses in IBD relative to healthcare worker (HCW) controls. Methods: We evaluated IBD patients enrolled in a US registry referred from 26 centers at 2, 8, and 16 weeks after completing 2 doses of SARSCoV- 2 mRNA vaccination and compared results to non-IBD non-immunosuppressed HCW participating in a parallel study. We analyzed plasma antibodies to the receptor binding domain of the viral spike protein using the SARS-CoV-2 IgG-II assay (Abbott Labs, Abbott Park, IL);IgG(S) > 50 AU/mL was defined as positive. Those with prior COVID were excluded. We also performed T-cell clonal analysis by T-cell receptor (TCR) immunosequencing at 8 weeks (Adaptive Biotechnologies, Seattle, WA). The breadth (number of unique sequences to a given protein) and depth (relative abundance of all the unique sequences to a given protein) of the T-cell clonal response were quantified using reference datasets. Analyses were adjusted for age, sex and vaccine type. Results: Overall, 1805 subjects were included (IBD n=1074 (65% Crohn's disease, 35% ulcerative colitis);HCW n=731). Age and sex were similar between both cohorts;Hispanic ethnicity and Asian race were less common among IBD than HCW (Table). Vaccine type included BNT162b2 (Pfizer) (75% of IBD, 98% of HCW) and the remainder mRNA-1274 (Moderna). IBD treatments included anti- TNF (46%), other biologics (33%), other immune suppressing therapy (9%), and no immune suppression (12%). Postvaccination antibody levels were lower among IBD than HCW both before and after adjusting for vaccine type (p<0.0001 each timepoint;Figure). After further restricting the IBD cohort to those on no immune-suppressive therapies, antibodies remained lower in IBD vs HCW at 2w (p=0.008) and 8w (p<0.0001), but not 16w (p=0.07). Among 321 subjects with available whole cell samples at 8 weeks (IBD n=163, HCW =158), Spikespecific TCR responses were similar between IBD and HCW for both clonal breadth and depth in both unadjusted and adjusted analyses;sub-analyses of those on biologics yielded similar results. Conclusion: Patients with IBD have dampened humoral responses, but similar cellular responses, after SARS-CoV-2 mRNA vaccination relative to HCW. These findings suggest a potentially greater risk of infection, but not of disease progression, among those with IBD, and should be considered to help guide booster dosing strategies for the IBD population. (Figure Presented) (Figure Presented) Figure: Post-vaccination immune responses: (A) Antibody responses are lower in IBD relative to non-IBD healthcare workers at 2, 8, and 16 weeks (p<0.0001 at each timepoint). In contrast, post-vaccination Spike-specific T-cell receptor clonal breadth (B1) and clonal depth (B2) at 8 weeks are similar in IBD compared to healthcare workers.

PARENTAL PERCEPTION OF COVID-19 VACCINATION IN CHILDREN UNDER 18 YEARS

Purpose of Study To identify factors influencing acceptance or refusal of COVID-19 vaccine by parents of children at a university affiliated community healthcare center. The introduction of the COVID-19 vaccine has been crucial in the mitigation of the COVID-19 pandemic. Initially, the vaccine was targeted at healthcare workers and other high-risk adults. Early in the pandemic, studies in adults on vaccine hesitancy noted concerns about the supposed novelty and efficacy of the vaccine. Studies in China and the UK established that most parents were willing to vaccinate their children with the COVID-19 vaccine even though they expressed concerns about safety and efficacy. This study aims to identify potential barriers and uncertainties that parents express when considering vaccinating their children. Methods Used Cross-sectional study using a self-administered questionnaire completed by English and Spanish-speaking parents/guardians requesting acute or well child care for their children less than 18 years of age at a university-affiliated pediatric ambulatory care center from April 2021 to May 2021. Summary of Results 223 subjects, predominantly mothers (90.1%), participated in the study. 49.8% of children were male and 48.4% female. The mean age of all children was 6.79 ± 5.4 years. 74% of the participants identified themselves as Hispanic, 17.9% as African American/Black, and 5.8% as 'mixed'. Concerning vaccinating their children against COVID-19, 23.8% of the respondents, stated 'Yes definitely' to agreeing to give the vaccine to their child, while 37.7% were 'Unsure but leaning towards yes'. 9.9% had 'No opinion', 13.5% were 'Unsure but leaning towards no', and 14.3% stated, 'No, definitely not'. Reasons for vaccine refusal include;uncertainty about vaccine efficacy (46.2%), concerns about hurried vaccine production (31.4%), belief that the child will get sick after vaccination (23.8%) and, being generally opposed to vaccines (4.5%). Conclusions In a largely Hispanic population, majority of parents/guardians were unsure but leaning towards accepting the COVID-19 vaccine for their children. Common reasons for vaccine refusal were concerns for efficacy and the rapid speed of production. This data suggests that specific vaccine education is needed in this community to address the concerns of efficacy, speedy vaccine production, and reactivity after vaccine administration to increase parental acceptance of COVID- 19 vaccine administration in their children.

Author Correction: Advancing COVID-19 diagnosis with privacy-preserving collaboration in artificial intelligence (Nature Machine Intelligence, (2021), 3, 12, (1081-1089), 10.1038/s42256-021-00421-z)

In the version of this article initially published, the first name of Chuansheng Zheng was misspelled as Chuangsheng. The error has been corrected in the HTML and PDF versions of the article. © The Author(s) 2022.

WORLDWIDE POST-MARKETING SAFETY SURVEILLANCE EXPERIENCE WITH TOFACITINIB IN ULCERATIVE COLITIS

Background: Tofacitinib is an oral, small molecule JAK inhibitor for the treatment of UC. First market authorization was received in the US in May 2018. Post-marketing surveillance (PMS) is an important part of monitoring adverse events (AEs). Here, we report an analysis of PMS case safety reports for tofacitinib in patients with UC.Methods: We analyzed the worldwide tofacitinib PMS reports received in the Pfizer safety database from May 30, 2018 to August 25, 2020. The type and estimated reporting rate (RR) of serious AEs (SAEs) of interest, incl. infection, vascular, respiratory, neoplasm, and cardiac events, were reviewed. Patient-years of exposure (PY) was estimated based on worldwide sales data and the calculated daily regimens of tofacitinib 5 or 10 mg twice daily, immediate or extended-release formulations.Results: During the 27-month reporting period, worldwide post-marketing exposure to tofacitinib was 8,916 PY. Overall, 4,226 case reports were received and included 12,103 AEs, of which 1,839 were SAEs. Among the cases reported, 1,141 (27.0%) included an SAE and 18 (0.4%) were fatal. Of cases with reported gender (88.1%) or age (81.6%), 46.5%occurred in men and the median age was 45 years (range 9–93). When analyzed by tofacitinib formulation, proportions of SAE cases were similar (Table 1). Table 2 presents a summary of AEs and SAEs by MedDRA system organ class. Among the 1,839 SAEs, RRs per 100 PY were 3.28 for infection events, 1.26 for vascular events, 0.74 for respiratory events, 0.55 for neoplasm events, and 0.50 for cardiac events. The most commonly reported serious infection events (MedDRA preferred term [PT] n≥8) were 2 PTs within the high level term (HLT) of Clostridia infections (C. difficile colitis/infection), pneumonia, COVID-19, cytomegalovirus, and herpes zoster. The most commonly reported serious vascular events (n≥10) included hemorrhage, thrombosis, and deep vein thrombosis. Most serious respiratory events were pulmonary embolism. The most commonly reported serious neoplasm events (n≥3) were 2 PTs within the HLT of breast and nipple neoplasms malignant (breast cancer female/breast cancer), colon cancer, lymphoma, malignant melanoma, neoplasm malignant, and prostate cancer. The most commonly reported serious cardiac events (n≥4) were 3 PTs within the HLT of ischemic coronary artery disorders (acute myocardial infarction/myocardial infarction, angina pectoris) and pericarditis.Conclusion: Based on this review of PMS data for tofacitinib in UC, the types of AEs and RRs were consistent with the known tofacitinib safety profile, with no new potential risks identified. Limitations of PMS reports, low numbers of case reports for extended-release formulation, and reliance on estimated RRs due to lack of precise values for exposure, required for incidence rate calculation, should be considered when interpreting these results.(Table Presented)(Table Presented)

Advancing COVID-19 diagnosis with privacy-preserving collaboration in artificial intelligence

Artificial intelligence provides a promising solution for streamlining COVID-19 diagnoses;however, concerns surrounding security and trustworthiness impede the collection of large-scale representative medical data, posing a considerable challenge for training a well-generalized model in clinical practices. To address this, we launch the Unified CT-COVID AI Diagnostic Initiative (UCADI), where the artificial intelligence (AI) model can be distributedly trained and independently executed at each host institution under a federated learning framework without data sharing. Here we show that our federated learning framework model considerably outperformed all of the local models (with a test sensitivity/specificity of 0.973/0.951 in China and 0.730/0.942 in the United Kingdom), achieving comparable performance with a panel of professional radiologists. We further evaluated the model on the hold-out (collected from another two hospitals without the federated learning framework) and heterogeneous (acquired with contrast materials) data, provided visual explanations for decisions made by the model, and analysed the trade-offs between the model performance and the communication costs in the federated training process. Our study is based on 9,573 chest computed tomography scans from 3,336 patients collected from 23 hospitals located in China and the United Kingdom. Collectively, our work advanced the prospects of utilizing federated learning for privacy-preserving AI in digital health. The COVID-19 pandemic sparked the need for international collaboration in using clinical data for rapid development of diagnosis and treatment methods. But the sensitive nature of medical data requires special care and ideally potentially sensitive data would not leave the organization which collected it. Xiang Bai and colleagues present a privacy-preserving AI framework for CT-based COVID-19 diagnosis and demonstrate it on data from 23 hospitals in China and the United Kingdom.

Efficacy of upadacitinib induction therapy in patients with moderately to severely active ulcerative colitis by biologic inadequate responder status: Results from two randomized phase 3 studies

Introduction: Upadacitinib (UPA), an oral JAK inhibitor, showed significantly greater efficacy vs placebo (PBO) in induction treatment of patients (pts) with moderately-to-severely active ulcerative colitis (UC) in two phase 3 induction trials, U-ACHIEVE and U-ACCOMPLISH. We evaluated efficacy of UPA in pts who had an inadequate response (IR), loss of response, or intolerance to biologic therapies (Bio-IR) or were non-Bio-IR. Methods: U-ACHIEVE and U-ACCOMPLISH, multicentre, double-blind, placebo (PBO)-controlled trials, randomized pts with moderately to severely active UC to UPA 45 mg QD or PBO for 8 weeks (wks). Randomization was stratified by status of previous biologic failure, ie an inadequate response (IR), loss of response, or intolerance to biologic therapies (Bio-IR or bio-failure) vs non-Bio-IR (nonbio-IR or non-bio-failure), baseline corticosteroid use (yes or no), and baseline adapted Mayo score (≤7 or>7). Efficacy endpoints included primary endpoint of clinical remission (adaptedMayo score) at Wk 8 and ranked secondary endpoints of clinical response (partial adapted Mayo score at Wk 2 and adapted Mayo score at Wk 8), endoscopic improvement (Mayo endoscopic subscore 0 or 1), endoscopic remission (Mayo endoscopic subscore 0) and histologic-endoscopic mucosal improvement at Wk 8 (HEMI;endoscopic subscore ≤1 and Geboes score ≤3.1). Results using non-responder imputation incorporating multiple imputation for missing data due to COVID-19 are reported. Results: In both studies, approximately half the pts were Bio-IR (Table 1). In both Bio-IR and non-Bio-IR pts, significantly higher proportion of pts receiving UPA achieved primary endpoint of clinical remission versus PBO;the magnitude of clinical remission at Wk 8 was greater in non-Bio-IR pts (UPA, 35% vs PBO, 9%;treatment difference [95% CI]: 26.0% [16.0, 36.1]) versus Bio-IR (UPA, 18% vs PBO, 0%;17.5% [11.4, 23.6]) in U-ACHIEVE and non-Bio-IR (UPA, 38% vs PBO, 6%;31.6% [22.8, 40.5]) versus Bio-IR (UPA, 30% vs PBO, 2%;27.1% [19.6, 34.7];Table 1) in U-ACCOMPLISH. Results were generally similar for ranked secondary endpoints (Table 1). UPA 45 mg QD was well-tolerated and no new safety signals were observed. Conclusion: UPA 45 mg QD is an effective induction treatment for pts with moderately to severely active UC. A significantly higher proportion of pts in both Bio-IR and non-Bio-IR groups receiving UPA achieved primary and secondary endpoints versus PBO. The magnitude of difference was greater among pts who were non-Bio-IR versus Bio-IR.

Extracting transmission and recovery parameters for an adaptive global system dynamics model of the COVID-19 pandemic

Accurately modelling the susceptibility, infection, and recovery of populations with regards to the COVID-19 pandemic is highly relevant for the implementation of countermeasures by governing bodies. In the past year, several thousands of articles on COVID-19 modelling were published. The spread of the pandemic has frequently been modelled using the Susceptible-Infected-Recovered (SIR) epidemic model owing to the low level of complexity. In recognition of its simplicity, we developed an SIR model to represent the spread of disease on a global scale, irrespective of mutation and countermeasures. The SIR parameters were reverse-engineered from aggregated global data. This model is the first to retrospectively deduce the initial incidence. The average transmission and recovery parameters were computed to be 0.33 week-1 and 0.23 week-1, respectively. These values lie well within the range of reported values on COVID-19 determined from geographically different regions. The model was simulated in the Ventana® simulation environment Vensim® for a 65-weeks duration and an adjusted initial infection incidence, which was presumed three times the reported initial infection incidence. The simulated data visually aligns with the real incidence data. We attribute the discrepancy between the presumed initial value and the reported value to lack of testing facilities on the starting date of 1 March 2020. Our parameter extraction suggests a novel methodology to quantify undertesting retrospectively in epidemics. © 2021 IEEE.

Volume balance and AKI in critically ill patients with SARS-CoV-2

Background: Evidence from the management of critically ill patients suggests restrictive volume management strategies and avoidance of volume overload improve ICU outcomes. Restrictive practices have been applied to the management of patients with SARS-CoV-2 (COVID-19), but data describing volume management and its associated outcomes in those with and without acute kidney injury (AKI) in this setting are lacking. Methods: We conducted a single-center retrospective cohort study of ICU patients with COVID-19. 7-day cumulative volume balance from ICU admission in excess of -5% (negative balance) or +5% (positive balance) of ICU admission weight as well as AKI based on KDIGO guidelines were identified. Associations between volume balance, AKI and clinical outcomes (dialysis, mechanical ventilation, and inpatient mortality) were explored. Results: 194 of 374 ICU admissions (51.9%) had AKI with 60 of 374 (194) (16.0%, 30.9% of those with AKI) requiring dialysis. 110 of 374 (29.4%) developed negative balance and 40 of 374 (10.7%) developed positive balance. Inpatient mortality was higher in those with AKI and negative balance (28%) and positive balance (35%) compared to those with neutral balance (16%) (p=0.039). Of the subjects with negative balance (Table), despite no difference in their net volume balance, inpatient mortality was significantly higher in those with AKI compared to those without AKI (p=0.048). Using the Kaplan-Meier estimator, patients with no AKI had no difference in inpatient survival when compared on the basis of volume balance (p=0.69). However, in those with AKI, inpatient survival was significantly lower for positive and negative volume balance compared to neutral balance (p=0.01). Conclusions: Negative and positive volume balance are associated with higher inpatient mortality particularly in patients with AKI. Future research must investigate the impact of negative balance on morbidity and mortality in patients with and without AKI.

SARS-CoV-2 infection in public school district employees following a district-wide vaccination program - Philadelphia county, Pennsylvania, March 21-April 23, 2021

The resumption of in-school instruction followed a mass vaccination program using the Pfizer-BioNTech 2-dose vaccine offered under a partnership between the Philadelphia Department of Public Health and Children's Hospital of Philadelphia to all 22,808 School District of Philadelphia employees during February 23-April 3, 2021. The subsequent mandatory testing program provided an opportunity to assess the percentage of positive BinaxNow point-of-care antigen tests (Abbott Laboratories) identified among school staff members based on their self-reported vaccination status at the time of testing. During the initial 5 weeks after schools reopened, 34,048 screening tests were performed. Overall, 0.70% of tests returned a positive result. The percentage of positive test results was lower among persons who reported receipt of 2 vaccine doses (0.09%) compared with those who reported receipt of 1 dose (1.21%) or zero doses (1.76%) representing a 95% reduction in percentage of positive SARS-CoV-2 test results among persons reporting receipt of 2 compared with zero doses of Pfizer-BioNTech vaccine. Vaccination of school staff members has been highlighted as an important strategy to maximize the safety of in-person education of K-12 students this fall. These findings reinforce the importance of promoting COVID19 vaccination among school staff members before commencement of the 2021-22 school year.

Role of cardiac biomarkers in cardiovascular manifestation of covid-19

Since early December 2019, an outbreak of pneumonia of unknown etiology has been reported in Wuhan, China The pathogen was then named Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) by WHO, and the illness caused by it was termed as the Coronavirus Disease 2019 (COVID-19). Currently, the disease has rapidly spread to the whole world and become an international public health emergency. Although the SARS novel coronavirus (SARS-CoV-2) or COVID-19 is a viral illness, in fact it is systemic illness in which most of the organ systems are affected with varying degree. There are various patterns of cardiovascular involvement in COVID 19. First, cardiovascular disease present as pre-existing comorbidity which becomes apparent or becomes more complicated and decompensated during COVID 19. Second, cardiovascular system involvement results due to systemic inflammatory response during the course COVID 19. Third, cardiovascular system can be affected during treatment due to the side effects of some medication or secondary hospital-acquired infections and complications. Arrhythmias, acute coronary syndrome (ACS), myocarditis, Heart Failure and cardiomyopathy are the most common cardiovascular disease in COVID-19. The Fourth Universal Definition of Myocardial Infarction defines myocardial injury (acute or chronic) as cTn concentrations >99th percentile upper reference limit (URL). In COVID-19, cardiac injury is believed to be through multiple overlapping factors such as severe inflammatory response with uncontrolled cytokine activation and/or direct injury due to Virus infiltration in cardiomyocytes through angiotensin-converting enzyme 2 (ACE 2) receptors. Increases in cardiac biomarkers, especially, cardiac troponin (cTn) are common in patients with COVID-19, particularly in patients with underlying cardiovascular conditions and severe COVID-19 presentations, and are associated with worse outcomes and mortality. Thus, it is evident that cardiac injury plays a significant role in the disease progression and outcome associated with COVID-19. Thus, it is reasonable to include the indicators of cardiac injury in the patient's diagnosis, triaging, treatment, and prognosis, while recognizing that their abnormality may not be related to direct coronary involvement. Incorporation of cardiac biomarkers measurement (cTn and/or B-type natriuretic peptide [BNP]) to a set of other inflammatory and thrombotic markers may facilitate the understanding of COVID-19 stages, risk profiles, and disease phenotypes. Baseline measurements can facilitate stage classification and initial triage, and serial measurements help with short- and long-term risk stratification (likelihood for survival and/or adverse events). This information is likely to be most beneficial in patients in whom disease stage and risk status is uncertain, as well as in patients in whom risk is particularly high. In both cases, cardiac biomarkers can help with decisions about COVID19 patients' triage, management, therapeutic treatment and level of care.

Patterns in cancer management changes for patients with COVID19 in northern California

Background: The COVID-19 pandemic affected oncology practice in ways that are still evolving. In particular, COVID-19 has led to changes in cancer treatment for patients (pts) infected with COVID, which may have long-term implications for both COVID and cancer-related outcomes. In this retrospective analysis, we describe changes in cancer management over time for cancer pts diagnosed with COVID-19 at two academic institutions in Northern California. Methods: Adult and pediatric pts diagnosed with COVID19 receiving active cancer management, defined as therapy/surgery/diagnostics within 3 weeks of COVID diagnosis, were identified through the EMR. Patients whose care was affected by COVID-19 were identified and analyzed for significant intra-group differences with regards to management type, treatment intent, and the time of COVID-19 diagnosis ('early' was defined as March to June 2020 and 'late' as July 2020 to January 2021). The duration and characteristics of such changes were compared across subgroups. Chi-squared test was used to compare the incidence of delays between subgroups. Results: Among 134 COVID-positive pts on active cancer management, 83 (62%) had significant changes in management that consisted primarily of treatment delays. More delays were identified in patients treated with curative intent earlier in the course of the pandemic compared to later (OR 4.1, p=0.022). This difference was not seen among pts treated with palliative intent. In addition, pts on oral (PO) therapy were significantly less likely to have treatment changes than those on IV/IM therapy (OR 0.32, p=0.005). This difference was driven by a decrease in management changes for those on PO therapy in the later time period (OR 0.27, p=0.026). Pts diagnosed later were more likely to have delays due to clinical reasons rather than institutional policy (OR 6.2, P<.005). The median delay in both time frames was 21 days. Comparison of subgroups is shown in the table. Conclusions: We found significant changes in management of cancer pts with COVID-19 that evolved over time. Oncologists have become increasingly willing to continue therapy for cancer pts treated with curative intent and pts on oral therapy. Changes in cancer therapy have become more frequently related to patient clinical status, and less so due to institutional policies. It will be important to analyze how these changes in management are ultimately reflected in cancer outcomes in order to equip patients and oncologists to react to the next pandemic.