ABSTRACT
Background: The evidence for management of severe COVID-19 with persistent respiratory failure (PRF) after acute treatment is scarce, despite some authors reported a good response to corticosteroid in histological proven secondary organising pneumonia (OP). Objective(s): This study aimed to study the disease course of COVID-19 patients with persistent respiratory failure, and its radiological pattern. Method(s): A single centre retrospective cohort study on severe COVID-19 patients was conducted from January 2021 to June 2021. All radiological imaging and data were retrieved from electronic database. Result(s): Severe COVID-19 pneumonia had a 78% (584/750) survival in our cohort. Among the survival, 48% (279/584) had PRF beyond 14 days of illness and 10% of them required oxygen therapy upon discharge. Eighty-six percent (240/279) of patients with PRF had a HRCT performed. Eighty percent (187/240) of them attended clinic follow up with 81% had a radiological pattern consistent with OP. The mean severity CT score was 10 (SD+/-3). [Jin C et al. Front Public Heal.2020;8] Seventy-eight percent of patients were perceived with WHO functional class of 1-2. Sixty-eight percent of patients (128/187) were given short course of prednisolone during admission with tapering doses. The mean prednisolone dose was 0.69mg/kg/day with a mean duration of 47 days (SD+/-18). Seventy-eight percent (146/187) had a follow up chest x-ray (CXR) at 12+/-8 weeks. Only 6.4% (12/187) of them had abnormal CXR findings whereby two patients were later confirmed to have pulmonary tuberculosis. Conclusion(s): Radiological pattern of OP is common in COVID-19 with PRF. HRCT is a non-invasive tool to assess this entity.
ABSTRACT
The COVID-19 pandemic has become a critical threat to global health and the economy since its first outbreak in 2019. The standard diagnosis for COVID-19, Reverse Transcription Polymerase Chain Reaction (RT-PCR) is time consuming, and has lower sensitivity compared to CT-scans. Therefore, CT-scans can be used as a complementary method, alongside RT-PCR tests for COVID-19 infection prediction. However, manually reviewing CT scans is time consuming. In this paper, we propose DECOVID-CT, a deep learning model based on 3D convolutional neural network (CNN) for the detection of COVID-19 infection with CT images. The model is trained and tested on the RICORD dataset, a multinational dataset, for higher robustness. Our model achieved an accuracy of 100%, for predicting COVID-19 positive images. © 2022 IEEE.
ABSTRACT
Background: We are a state hospital in Selangor, managing 180,000-200,000 patients per year. During the COVID-19 pandemic, we served as a hybrid hospital, managing both COVID and non-COVID patients. When Malaysia was hit by a third wave of COVID-19 in June 2021, our department was badly affected. The number of COVID-19 patients needing hospital treatment increased exponentially. Our local disaster alert reached level 5 within 2 weeks. Objectives: There are multiple methods and approach to manage the disaster, and no one was better than the other. Here, we are sharing our responses and approaches in managing COVID-19 disaster. Method: Our approach was a form of integrated clinical support to control the upstream, maintaining a good midstream and an effective downstream of patients. It was a team that consisted of clinician, bed resource manager, ambulance and transport manager, and logistic manager. The team caters to the three important resource elements by creating and optimizing the space available, increasing the number of staff, and maintaining adequate supply for clinical needs. Result: With an integrated team, our disaster response was more organized and focused. We managed to increase the space and optimize the staff capacity. Our baseline bed waiting time (BWT) in May and June 2021 was between 16 and 20 h. We reached a maximum of 110 h in mid-July. With aggressive collaboration from different teams, we managed to reduce BWT by 50% within 2 weeks and back to baseline in 4-week duration. Supplies were continuously available and sufficient by simplifying the bureaucracy of item procurement and donation policy. Conclusion: An unprecedented situation demands an extraordinary response. Our strategy in integrated support has shown a good response in managing resources during disaster. A team with flexibility, creativity, and innovation is the key to success.
ABSTRACT
There is increased infection risk at the time of autologous stem cell transplantation (ASCT) including for patients with plasma cell disorders (PCD), therefore preventing infection with COVID-19 vaccination in this vulnerable group is key. However, patients with PCD have been shown to mount suboptimal responses to COVID-19 vaccination. A clinical audit of serological response to COVID-19 vaccination before and after ASCT was undertaken, to observe how antibody titres change during this period. Antibodies to the SARS-CoV-2 spike protein were measured using the Elecsys Anti-SARS-CoV-2S assay (Roche diagnostics) in 88 patients who underwent ASCT for PCD at the University College London Hospital NHS Foundation Trust between December 2020 and September 2021. Pre-ASCT antibody titres were measured following first or second vaccine and following ASCT. The majority ( n = 76) had no prior history of COVID-19 infection, and four of this cohort declined vaccination. In those who received one vaccine pre-ASCT ( n = 21), 76% seroconverted with a median titre of 11.3 3 U/ml (IQR 1.5-62.6). In those who received two doses pre-ASCT ( n = 51), 97% seroconverted with a median titre of 494 U/ml (IQR 190.5-1681). In those who received two doses pre-ASCT, anti-S antibodies were detected in the immediate post-ASCT setting, with titres of 373 U/ml (median, IQR 40.6-2326) measured less than or equal to 28 days (median 15 [6-25]) post-ASCT, and 170 U/ml (IQR 55-604) at more than 28 days (median 85 [32-125]) post-ASCT. Patients who received one dose pre-ASCT had lower median titres of 36.5 U/ml (IQR 12.6-1310) measured less than or equal to 28 days (median 15 [12-22] post-ASCT and 7.7 U/ml (IQR 2.9-23.8) at more than 28 days (median 85 [40-104] post-ASCT. Antibody levels declined over time, but patients who had received two vaccines pre-ASCT maintained higher titres post-ASCT compared to those who had received one dose, emphasising the importance of COVID-19 vaccination prior to ASCT. Our patients are advised to be re-vaccinated against COVID-19 3 months after ASCT, and antibody response following re-vaccination was measured in a subgroup ( n = 14). Those who were previously un-vaccinated did not seroconvert following one dose. However, antibody titres in those who had received either one or two vaccines ( n = 12) prior to ASCT increased from 32.4 U/ml (median, IQR 13.4-1082) post-ASCT to 431 U/ml (median, IQR 15.33-2500) following re-vaccination. Those who had received two vaccines pre-ASCT ( n = 2) achieved higher titres than those who had received a single dose. In conclusion, we demonstrated how protective titres fall during the patient's journey through ASCT and our repeated interactions with them. Despite this, patients vaccinated prior to ASCT maintain some level of measurable antibody immediately post-ASCT, which is encouraging as patients are considered most vulnerable to infection during this period. Titres were also boosted effectively after one dose of re-vaccination, compared to those never vaccinated. Current guidance is for adult patients who have undergone ASCT to be considered 'never vaccinated' against COVID-19, in line with pre-COVID-19 re-vaccination practice, and to receive a three-dose primary course followed by a booster vaccination post-ASCT. We must facilitate and encourage our patients to be vaccinated prior and after ASCT in this rapidly changing landscape, especially in context of the spread and evolution of a potentially more transmissible virus. (Table Presented).
ABSTRACT
Patients with haematological malignancies, such as acute leukaemia (AL) and high-risk MDS (HRMDS), have significantly increased mortality and morbidity from COVID-19 but vaccine efficacy in these patients remains to be fully established. To date, seroconversion rates following SARS-CoV-2 vaccination in patients with AL and HR-MDS have been reported in small numbers within large retrospective studies, with most not receiving active treatment We characterised the serological responses of patients with AL and HR-MDS who are receiving anti-cancer therapy and were vaccinated within the UK vaccination and programme, receiving two doses of either BNT162b2 or ChAdOx1nCoV-19. Fifty-five patients (35 AML [64%], 14 ALL [25%], 6 HRMDS [11%]), underwent serological testing for anti-S antibody levels after receiving 2 SARS-CoV-2 vaccine doses (71% BNT162b2, 16% ChAdOx1nCoV-19, 13%unknown), between December 2020 and July 2021, Table 1. Serological testing was performed per clinical practice using the Roche Elecsys anti-SARS-CoV-2 S enzyme immunoassay (seropositive results being 0.8 U/ml and upper limit >2500 U/ml). The median age was 53 (range 18-76) and 53 patients (96%) were receiving SACT. Serological testing for anti-N antibodies performed on 48 patients (87%) identified 11 (20%) seropositive patients, indicative of previous natural infection. Fifty (91%) were seropositive for anti-S antibodies after two doses (median 43 days [range 4-133] post dose), of these, 33 (60%) were tested following dose 1 (median 36 days [range 24-86]). 76% were seropositive after dose 1, with median antibody titres 7.88 U/ml (IQR 0.7-2.37). This rose to 91% following dose 2, with median 333 U/ml (IQR 27.5-1821). Higher rates of seropositivity are observed in AML/HR-MDS patients (95%) compared to ALL (79%). Patients with AML/HR-MDS also showed a significant increased seropositivity (95%) and titres following dose 2 (median 352 U/ml [IQR 89.95-2116], p = 0.0003), compared to patients with ALL who showed no increase (median 8.395 [IQR 2.335-669.3]). To precisely define seroconversion rates, we excluded patients with anti-N antibodies (indicative of previous SARSCoV-2 infection). SARS-CoV-2 naïve patients with AML/ HR-MDS had higher seroconversion rates and median anti-S antibody titres compared to ALL (median 291 U/ml [IQR 87.03-1771] vs. 5.06 U/ml [1.69-175.6], p = 0.005) and also showed significant increases in titres between post dose 1 and 2 dose titres, not seen in ALL. Patients with previous SARS-CoV-2 infection (anti-N positive) had higher antibody titres, following vaccination in AML/MDS (median 2500 U/ ml [IQR 141-2500]) compared to ALL (median 1541 U/ml [IQR 574.5-2500]), although not statistically significant. There was no significant difference in anti-S antibody response to vaccination in patients treated with intensive or non-intensive AML therapy but significantly reduced anti-S antibody titres were present in patients who received venetoclax-based regimens compared to other therapies ( n = 20, median 158.5 U/ml [IQR 34.85-873] vs. n = 19 median 796 U/ml [IQR 132-2500] p = 0.04). Patients with AML and HR-MDS on SACT are able to generate robust serological responses to SARS-CoV-2 infection but this is not the case for all patients following vaccination. Understanding impact of disease subtypes and therapy on vaccine response is pertinent, as decisions on modifying or delaying treatment in the context of either SARS-CoV-2 infection or vaccination require a clear evidence base. (Table Presented).
ABSTRACT
Background Plasma cell disorders (PCD) are at risk of inadequate immune responses to COVID-19 vaccines due to recognised humoral and cellular immune dysfunction which is multi-factorial and related to host and disease factors. With an estimated risk of 33% mortality from contracting COVID-19 in this population, protection with an anti-SARS-CoV-2 vaccination is critical. Initial extension to vaccination intervals in the United Kingdom to 12 weeks in December 2020 led to concerns that PCD patients would be left vulnerable for an extended period. Methods A clinical audit was performed on measured serological responses in PCD patients after first and second doses of the BNT162b2 and ChAdOx-1 nCoV-19 vaccines. Antibody levels were measured using Elecsys Anti-SARS-CoV-2S assay (Roche) for quantitative detection of IgG Abs, specific for the SARS-CoV-2 spike-protein. Positive cut-off of 0.80 U/mL defined serological response. Testing was performed at (or closest to) 4 and 8-weeks post-dose. Baseline nucleocapsid Ab results were available from previous screening in a subset of patients. All patients on CIT underwent 4-weekly swabs. Clinical information was retrieved from medical records. Results 188 PCD patients (155 multiple myeloma, 18 amyloid, 10 SMM/MGUS, other 5 PCD), median age 64 (range 32-84), had serological assessment after both vaccine doses. Fourteen with previous COVID-19 infection were excluded. Of 174 patients, 112 were tested after first dose. 88% (153) were on chemo-immunotherapy treatment (CIT). Seropositive rate after first dose was 63% (71/112);of those with available negative baseline antibody test, 62% (31/50) seroconverted. After second dose, 89% (154/174) were seropositive;of those with negative baseline antibody, 90% (61/68) seroconverted. Expectedly, paired median titres after second dose were significantly higher than post first dose (n=112, 3.245 U/mL (IQR 0.4-25.55) vs 518 U/mL (IQR 29.40-2187) p<0.0001) (Figure 1A). Of 41 patients seronegative after first dose, 25 (61%) seroconverted after second, though with lower titres than those only requiring one dose (Figure 1B). Active CIT, disease response less than PR, >=4 lines therapy, light-chain disease, male gender and not responding to first dose were significant factors for not responding to two vaccine doses. We explored <400 U/mL as sub-optimal response (in keeping with upcoming booster study eligibility, OCTAVE-DUO(1), also encompassing the lower quartile of reported healthy controls(2)), which included 43% (75/174) patients. Age 70 years, male gender, >=4 lines of treatment were independent predictors of less-than-optimal response (anti-CD38 CIT of borderline significance). Importantly, vaccine dosing intervals classified as =<42 vs >42 days (Figure 1C) or 28 +/- 14 days vs 84 +/- 14 days (excluding n=66 in neither) (Figure 1D) did not show difference in both definitions of response, neither did vaccine type. Fourteen with previous COVID-19 infection responded to one vaccine dose, median titres 2121 U/mL (IQR 23.48- 2500)) rising to median 2500 U/mL (IQR 2500-2500) after second dose (Figure 1E), significantly higher than those without previous infection. Conclusion Serological response to COVID-19 vaccine is lower in PCD patients than reported healthy controls at 63% after first dose, rising to 89% after second dose, despite extended dosing intervals. PCD patients should be prioritised for shorter intervals, as we show that patients seronegative after first dose, respond after second dose. Further work in PCD is needed to understand how Ab levels correlate to neutralisation capability, cellular responses, protection from infection and how long seroconversion lasts to better define correlates of protection. A booster vaccination or prophylactic passive antibody strategy may be required for those identified at risk, shown not to have responded to two vaccine doses or to have less-than-optimal response. Results from these trials will be eagerly awaited. (Figure Presented).
ABSTRACT
The Coronavirus Disease (COVID-19) has caused millions of casualties across the globe. One inexpensive and noninvasive screening method for COVID-19 is the analysis of chest X-ray (CXR) images for pathological features in the lungs. These features are difficult to detect by humans, but convolutional neural networks (CNN) have proven effective at extracting them. This paper uses four ImageNet-pre-trained CNNs: VGG16, DenseNet201, ResNet50, and EfficientNetB3 to perform transfer learning to a task of COVID-19 CXR image detection on a dataset containing COVID-19, healthy, and viral pneumonia CXR images. We compare the performance of the retrained CNNs using standard measures and investigate the features they use for their predictions using local interpretable model-agnostic explanations (LIME). The networks are retrained on two classification tasks: Task 1 consists of classifying healthy and COVID-19 CXR images and task 2 consists of classifying viral pneumonia and COVID-19 CXR images. We find that DenseNet201 and VGG16 achieve higher accuracies than ResNet50 and EfficientN etB3 in both tasks. However, the LIME explanations reveal that VGG16 does not learn disease-relevant features in the lungs, while DenseNet201, ResNet50, and EfficientNetB3 use regions in the lungs to make their predictions. This observation is reinforced by comparing LIME explanations with ground-truth lung regions on an unseen dataset. The prospect of using 'black box' deep neural networks for automatic screening of CXRs for COVID-19 can be improved with LIME-enabled investigations of model performance. © 2021 IEEE.
ABSTRACT
Background: Plasma cell disorder (PCD) patients are extremely vulnerable to SARS-CoV-2 infection due to disease-related impaired humoral and cellular immunity as well as the receipt of immunosuppressive therapy. Reported mortality in a large cohort of patients with plasma cell disorders is 33%. The roll out of the COVID-19 vaccine is welcomed in this population, however there is concern of suboptimal antibody responses, from previous experience with the influenza vaccine. There is urgent need to understand the humoral response to SARSCoV- 2 infection in these patients, in the context of systemic anti-cancer therapy (SACT). Aims: We aimed to investigate the presence of SARS-CoV-2 antibodies in a cohort of PCD patients, the relationship with symptomatic infection, PCD characteristics and receipt of SACT. Methods: SARS-CoV-2 antibody screening with the Elecsys Anti-SARSCoV- 2 assay (Roche Diagnostics, Basel, Switzerland), a semi-quantitative assay of IgG and IgM against the nucleocapsid (N) antigen was introduced for PCD patients at our institution in July 2020. Clinical information was retrieved from the medical records. Patients with unexpected positive antibody tests were asked about possible past contacts and exposure to SARS-CoV-2. Results: We report on a six-month period of routine SARS-CoV-2 antibody screening. Two-hundred and forty-three PCD patients had one antibody test, 106 had serial samples. Total seroprevalence was 10.7% (26/243), of which 12 were patients with known PCR-swab positive COVID-19 disease. In a separate but overlapping cohort, 41 patients have had PCR confirmed COVID-19 disease;20 of these patients were tested, and 12 (60%) had seroconverted. Median time to testing from positive PCR test in the antibody positive patients was 86.5 days (range 22-256) and in antibody negative patients, 30.5 days (range 5-176 days). No PCD or COVID-19 disease factors were found to influence the likelihood of mounting an antibody response after PCR-confirmed COVID-19 disease in these 20 patients. In our screened cohort, 14 (6.3%) patients were unexpectedly antibody positive. Their clinical course is summarised in the included figure. The majority 85.7% (12/14) of patients described no COVID-19 symptoms. Seven (50%) patients were on SACT (including ixazomib, pomalidomide, lenalidomide and dexamethasone combinations) throughout the period from possible exposure to positive antibody test, with no interruption to their ongoing oral immunomodulatory treatment. Ten antibody positive patients had serial positive results at median 45 days (range 21-119) apart, demonstrating persistence, but some decline in titre over time. Summary/Conclusion: Our seroprevalence of 10.7% is lower but not dissimilar to that reported in the London population over a similar time period reflecting shielding behaviours in our patients but also the challenges of protecting them during high SARS-CoV-2 incidence in the community. Nevertheless, PCD patients retain the ability to seroconvert, even with asymptomatic COVID-19 disease and while on immunomodulatory therapy. Seroconversion rates following symptomatic infection appear lower however, with evidence of delay compared to the general population. These data support the advice for COVID-19 vaccination to be offered to all PCD patients although the suboptimal humoral response calls for close antibody monitoring of all vaccinated PCD patients and timely booster doses.